Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252

27 posts in General Discussion Last posting was on 2004-03-25 12:30:54.0Z
Dave Savage Posted on 2004-02-13 10:25:21.0Z
Sender: 647d.402ca5a4.1804289383@sybase.com
From: Dave Savage
Newsgroups: ianywhere.public.general
Subject: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <402ca611.6481.846930886@sybase.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 13 Feb 2004 02:25:21 -0800
X-Trace: forums-1-dub 1076667921 10.22.241.41 (13 Feb 2004 02:25:21 -0800)
X-Original-Trace: 13 Feb 2004 02:25:21 -0800, 10.22.241.41
Lines: 16
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2479
Article PK: 5619

Working with Sybase SQL Anywhere Verion 9.0.0. Build 1108

When Two Threads created a DeadLock, I get the Correct
Deadlock Error returned for one of the Threads,

However when I install the Express Bug Fix, Update of 9.0.0
to Build 1252

I no longer get the DeadLock Error, The ASA Commands just
Hang.

Has anyone else seen this ?

Thanks in advance

Dave


Breck Carter [TeamSybase] Posted on 2004-02-13 13:40:16.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com>
References: <402ca611.6481.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: bcarter.sentex.ca
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
Date: 13 Feb 2004 05:40:16 -0800
X-Trace: forums-1-dub 1076679616 64.7.134.118 (13 Feb 2004 05:40:16 -0800)
X-Original-Trace: 13 Feb 2004 05:40:16 -0800, bcarter.sentex.ca
Lines: 31
X-Authenticated-User: TeamPS
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2481
Article PK: 5618

FWIW I just ran a deadlock test using two connections and got
"Deadlock detected" as expected, in build 1252.

What does sa_conn_info show when you run it on a 3rd connection? Does
it show an actual lock cycle (deadlock)?

Breck

On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:

>Working with Sybase SQL Anywhere Verion 9.0.0. Build 1108
>
>When Two Threads created a DeadLock, I get the Correct
>Deadlock Error returned for one of the Threads,
>
>However when I install the Express Bug Fix, Update of 9.0.0
>to Build 1252
>
>I no longer get the DeadLock Error, The ASA Commands just
>Hang.
>
>Has anyone else seen this ?
>
>Thanks in advance
>
>Dave

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


Dave Savage Posted on 2004-02-13 16:26:34.0Z
Sender: 69ba.402cf8aa.1804289383@sybase.com
From: Dave Savage
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <402cfaba.69da.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 13 Feb 2004 08:26:34 -0800
X-Trace: forums-1-dub 1076689594 10.22.241.41 (13 Feb 2004 08:26:34 -0800)
X-Original-Trace: 13 Feb 2004 08:26:34 -0800, 10.22.241.41
Lines: 41
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2488
Article PK: 5626

I found the "Deadlock Detected" worked when the threads were
in separate applications but not when in the same
application

It did show a lock Cycle, Only 3 out of the 4 Updates to the
database were processed by the server.

Dave

> FWIW I just ran a deadlock test using two connections and
> got "Deadlock detected" as expected, in build 1252.
>
> What does sa_conn_info show when you run it on a 3rd
> connection? Does it show an actual lock cycle (deadlock)?
>
> Breck
>
> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>
> >Working with Sybase SQL Anywhere Verion 9.0.0. Build 1108
> >
> >When Two Threads created a DeadLock, I get the Correct
> >Deadlock Error returned for one of the Threads,
> >
> >However when I install the Express Bug Fix, Update of
> 9.0.0 >to Build 1252
> >
> >I no longer get the DeadLock Error, The ASA Commands just
> >Hang.
> >
> >Has anyone else seen this ?
> >
> >Thanks in advance
> >
> >Dave
>
> --
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-13 16:52:54.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
X-Original-Trace: 13 Feb 2004 09:03:59 -0800, bcarter.sentex.ca
Lines: 57
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 08:34:46 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 08:52:54 -0800
X-Trace: forums-1-dub 1076691174 10.22.108.75 (13 Feb 2004 08:52:54 -0800)
X-Original-Trace: 13 Feb 2004 08:52:54 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2489
Article PK: 5627

So sa_conn_info showed two blocked connections, blocked on each other?
That's a situation that should never actually be viewable IMHO.

Are you sure you're on 1252, not build 1270?

Please excuse my dumb questions, but I am chasing a related issue.

Breck

On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:

>I found the "Deadlock Detected" worked when the threads were
>in separate applications but not when in the same
>application
>
>It did show a lock Cycle, Only 3 out of the 4 Updates to the
>database were processed by the server.
>
>Dave
>
>
>> FWIW I just ran a deadlock test using two connections and
>> got "Deadlock detected" as expected, in build 1252.
>>
>> What does sa_conn_info show when you run it on a 3rd
>> connection? Does it show an actual lock cycle (deadlock)?
>>
>> Breck
>>
>> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>>
>> >Working with Sybase SQL Anywhere Verion 9.0.0. Build 1108
>> >
>> >When Two Threads created a DeadLock, I get the Correct
>> >Deadlock Error returned for one of the Threads,
>> >
>> >However when I install the Express Bug Fix, Update of
>> 9.0.0 >to Build 1252
>> >
>> >I no longer get the DeadLock Error, The ASA Commands just
>> >Hang.
>> >
>> >Has anyone else seen this ?
>> >
>> >Thanks in advance
>> >
>> >Dave
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


modan Posted on 2004-02-13 17:12:18.0Z
Sender: 68ac.402cffef.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <402d0802.6961.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com>
X-Original-NNTP-Posting-Host: 10.22.241.42
X-Original-Trace: 13 Feb 2004 09:23:14 -0800, 10.22.241.42
Lines: 73
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 08:54:01 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 09:12:18 -0800
X-Trace: forums-1-dub 1076692338 10.22.108.75 (13 Feb 2004 09:12:18 -0800)
X-Original-Trace: 13 Feb 2004 09:12:18 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2494
Article PK: 5632

Hi, I'm going to be taking this up, as Dave is off on
holiday.

Anyway, I have confirmed that the build is 1252.

In answer to your other question, sa_conn_info shows one
blocked connection, not two.

Both show ReqType of STMT_EXECUTE, and identical
lastreqtime, and UncmtOps = 0. The "unblocked" one shows a
LastIdle greater than the blocked one by either 1 or 2.

Modan

> So sa_conn_info showed two blocked connections, blocked on
> each other? That's a situation that should never actually
> be viewable IMHO.
>
> Are you sure you're on 1252, not build 1270?
>
> Please excuse my dumb questions, but I am chasing a
> related issue.
>
> Breck
>
> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>
> >I found the "Deadlock Detected" worked when the threads
> were >in separate applications but not when in the same
> >application
> >
> >It did show a lock Cycle, Only 3 out of the 4 Updates to
> the >database were processed by the server.
> >
> >Dave
> >
> >
> >> FWIW I just ran a deadlock test using two connections
> and >> got "Deadlock detected" as expected, in build 1252.
> >>
> >> What does sa_conn_info show when you run it on a 3rd
> >> connection? Does it show an actual lock cycle
> (deadlock)? >>
> >> Breck
> >>
> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> >>
> >> >Working with Sybase SQL Anywhere Verion 9.0.0. Build
> 1108 >> >
> >> >When Two Threads created a DeadLock, I get the Correct
> >> >Deadlock Error returned for one of the Threads,
> >> >
> >> >However when I install the Express Bug Fix, Update of
> >> 9.0.0 >to Build 1252
> >> >
> >> >I no longer get the DeadLock Error, The ASA Commands
> just >> >Hang.
> >> >
> >> >Has anyone else seen this ?
> >> >
> >> >Thanks in advance
> >> >
> >> >Dave
> >>
> >> --
> >> bcarter@risingroad.com
> >> Mobile and Distributed Enterprise Database Applications
> >> www.risingroad.com
>
> --
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-13 22:25:13.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <9nhq20ltmvodjuvdd61v78hrrc3s943dg4@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
X-Original-Trace: 13 Feb 2004 14:36:17 -0800, bcarter.sentex.ca
Lines: 109
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 14:07:02 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 14:25:13 -0800
X-Trace: forums-1-dub 1076711113 10.22.108.75 (13 Feb 2004 14:25:13 -0800)
X-Original-Trace: 13 Feb 2004 14:25:13 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2498
Article PK: 5634

I just ran a series of tests on builds 1250 and 1252 and got identical
results... a deadlock is instantly detected and resolved by undoing
the atomic operation that caused the deadlock and raising SQLSTATE
40001 on that connection.

One interesting thing I did discover is the following: the SQL
Anywhere engine will automatically issue a ROLLBACK if any request
from the client interface finishes with SQLSTATE set to 40001
(deadlock detected). This means that a SIGNAL statement can cause a
ROLLBACK, for example, as well as an operation that causes a real
deadlock.

On the other hand, if an operation that sets SQLSTATE to 40001 is
inside a BEGIN block with an EXCEPTION handler that catches the error
and doesn't RESIGNAL it, the SQLSTATE will no longer be 40001 when the
client request is finished, and the automatic ROLLBACK will not be
performed. In this case, earlier locks obtained by the connection that
got the 40001 error will *still* be held, although no new locks will
be obtained by the failed operation. This means the other connection
in the deadlock pair will still be blocked, but the connection that
got the error is free to proceed.

This is old behaviour and it hasn't changed, and I'll bet it doesn't
explain your observation.

...unless you changed your code between 1250 and 1252 :)

Breck The Deadlocker

On 13 Feb 2004 09:12:18 -0800, modan wrote:

>Hi, I'm going to be taking this up, as Dave is off on
>holiday.
>
>Anyway, I have confirmed that the build is 1252.
>
>In answer to your other question, sa_conn_info shows one
>blocked connection, not two.
>
>Both show ReqType of STMT_EXECUTE, and identical
>lastreqtime, and UncmtOps = 0. The "unblocked" one shows a
>LastIdle greater than the blocked one by either 1 or 2.
>
>Modan
>
>> So sa_conn_info showed two blocked connections, blocked on
>> each other? That's a situation that should never actually
>> be viewable IMHO.
>>
>> Are you sure you're on 1252, not build 1270?
>>
>> Please excuse my dumb questions, but I am chasing a
>> related issue.
>>
>> Breck
>>
>> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>>
>> >I found the "Deadlock Detected" worked when the threads
>> were >in separate applications but not when in the same
>> >application
>> >
>> >It did show a lock Cycle, Only 3 out of the 4 Updates to
>> the >database were processed by the server.
>> >
>> >Dave
>> >
>> >
>> >> FWIW I just ran a deadlock test using two connections
>> and >> got "Deadlock detected" as expected, in build 1252.
>> >>
>> >> What does sa_conn_info show when you run it on a 3rd
>> >> connection? Does it show an actual lock cycle
>> (deadlock)? >>
>> >> Breck
>> >>
>> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>> >>
>> >> >Working with Sybase SQL Anywhere Verion 9.0.0. Build
>> 1108 >> >
>> >> >When Two Threads created a DeadLock, I get the Correct
>> >> >Deadlock Error returned for one of the Threads,
>> >> >
>> >> >However when I install the Express Bug Fix, Update of
>> >> 9.0.0 >to Build 1252
>> >> >
>> >> >I no longer get the DeadLock Error, The ASA Commands
>> just >> >Hang.
>> >> >
>> >> >Has anyone else seen this ?
>> >> >
>> >> >Thanks in advance
>> >> >
>> >> >Dave
>> >>
>> >> --
>> >> bcarter@risingroad.com
>> >> Mobile and Distributed Enterprise Database Applications
>> >> www.risingroad.com
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-13 19:57:15.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
X-Original-Trace: 13 Feb 2004 12:08:19 -0800, bcarter.sentex.ca
Lines: 89
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 11:39:05 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 11:57:15 -0800
X-Trace: forums-1-dub 1076702235 10.22.108.75 (13 Feb 2004 11:57:15 -0800)
X-Original-Trace: 13 Feb 2004 11:57:15 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2496
Article PK: 5635

What settings do you use for the BLOCKING and BLOCKING_TIMEOUT
options?

Is it possible that the difference between 1250 and 1252 is that a
different connection is being chosen as the "victim" when a deadlock
is detected?

Breck

On 13 Feb 2004 09:12:18 -0800, modan wrote:

>Hi, I'm going to be taking this up, as Dave is off on
>holiday.
>
>Anyway, I have confirmed that the build is 1252.
>
>In answer to your other question, sa_conn_info shows one
>blocked connection, not two.
>
>Both show ReqType of STMT_EXECUTE, and identical
>lastreqtime, and UncmtOps = 0. The "unblocked" one shows a
>LastIdle greater than the blocked one by either 1 or 2.
>
>Modan
>
>> So sa_conn_info showed two blocked connections, blocked on
>> each other? That's a situation that should never actually
>> be viewable IMHO.
>>
>> Are you sure you're on 1252, not build 1270?
>>
>> Please excuse my dumb questions, but I am chasing a
>> related issue.
>>
>> Breck
>>
>> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>>
>> >I found the "Deadlock Detected" worked when the threads
>> were >in separate applications but not when in the same
>> >application
>> >
>> >It did show a lock Cycle, Only 3 out of the 4 Updates to
>> the >database were processed by the server.
>> >
>> >Dave
>> >
>> >
>> >> FWIW I just ran a deadlock test using two connections
>> and >> got "Deadlock detected" as expected, in build 1252.
>> >>
>> >> What does sa_conn_info show when you run it on a 3rd
>> >> connection? Does it show an actual lock cycle
>> (deadlock)? >>
>> >> Breck
>> >>
>> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>> >>
>> >> >Working with Sybase SQL Anywhere Verion 9.0.0. Build
>> 1108 >> >
>> >> >When Two Threads created a DeadLock, I get the Correct
>> >> >Deadlock Error returned for one of the Threads,
>> >> >
>> >> >However when I install the Express Bug Fix, Update of
>> >> 9.0.0 >to Build 1252
>> >> >
>> >> >I no longer get the DeadLock Error, The ASA Commands
>> just >> >Hang.
>> >> >
>> >> >Has anyone else seen this ?
>> >> >
>> >> >Thanks in advance
>> >> >
>> >> >Dave
>> >>
>> >> --
>> >> bcarter@risingroad.com
>> >> Mobile and Distributed Enterprise Database Applications
>> >> www.risingroad.com
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


modan Posted on 2004-02-16 12:10:16.0Z
Sender: 2347.4030b30e.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <4030b328.234a.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 16 Feb 2004 04:10:16 -0800
X-Trace: forums-1-dub 1076933416 10.22.241.41 (16 Feb 2004 04:10:16 -0800)
X-Original-Trace: 16 Feb 2004 04:10:16 -0800, 10.22.241.41
Lines: 96
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2510
Article PK: 5647

Just to be clear, our program invokes 2 threads. The same
program can be run in a different mode where the program
only runs one of the threads, allowing us to run two
instances. When we run in this mode (two applications
rather than two threads) then the deadlock detection works,
and behaves as expected.

> What settings do you use for the BLOCKING and
> BLOCKING_TIMEOUT options?
>
> Is it possible that the difference between 1250 and 1252
> is that a different connection is being chosen as the
> "victim" when a deadlock is detected?
>
> Breck
>
> On 13 Feb 2004 09:12:18 -0800, modan wrote:
>
> >Hi, I'm going to be taking this up, as Dave is off on
> >holiday.
> >
> >Anyway, I have confirmed that the build is 1252.
> >
> >In answer to your other question, sa_conn_info shows one
> >blocked connection, not two.
> >
> >Both show ReqType of STMT_EXECUTE, and identical
> >lastreqtime, and UncmtOps = 0. The "unblocked" one shows
> a >LastIdle greater than the blocked one by either 1 or 2.
> >
> >Modan
> >
> >> So sa_conn_info showed two blocked connections, blocked
> on >> each other? That's a situation that should never
> actually >> be viewable IMHO.
> >>
> >> Are you sure you're on 1252, not build 1270?
> >>
> >> Please excuse my dumb questions, but I am chasing a
> >> related issue.
> >>
> >> Breck
> >>
> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> >>
> >> >I found the "Deadlock Detected" worked when the
> threads >> were >in separate applications but not when in
> the same >> >application
> >> >
> >> >It did show a lock Cycle, Only 3 out of the 4 Updates
> to >> the >database were processed by the server.
> >> >
> >> >Dave
> >> >
> >> >
> >> >> FWIW I just ran a deadlock test using two
> connections >> and >> got "Deadlock detected" as expected,
> in build 1252. >> >>
> >> >> What does sa_conn_info show when you run it on a 3rd
> >> >> connection? Does it show an actual lock cycle
> >> (deadlock)? >>
> >> >> Breck
> >> >>
> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> >> >>
> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> Build >> 1108 >> >
> >> >> >When Two Threads created a DeadLock, I get the
> Correct >> >> >Deadlock Error returned for one of the
> Threads, >> >> >
> >> >> >However when I install the Express Bug Fix, Update
> of >> >> 9.0.0 >to Build 1252
> >> >> >
> >> >> >I no longer get the DeadLock Error, The ASA
> Commands >> just >> >Hang.
> >> >> >
> >> >> >Has anyone else seen this ?
> >> >> >
> >> >> >Thanks in advance
> >> >> >
> >> >> >Dave
> >> >>
> >> >> --
> >> >> bcarter@risingroad.com
> >> >> Mobile and Distributed Enterprise Database
> Applications >> >> www.risingroad.com
> >>
> >> --
> >> bcarter@risingroad.com
> >> Mobile and Distributed Enterprise Database Applications
> >> www.risingroad.com
>
> --
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Robert Waywell Posted on 2004-02-16 15:15:19.0Z
From: "Robert Waywell" <nospam_rwaywell@ianywhere.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 126
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
NNTP-Posting-Host: rwaywell-pc.sybase.com
X-Original-NNTP-Posting-Host: rwaywell-pc.sybase.com
Message-ID: <4030de87$1@forums-1-dub>
Date: 16 Feb 2004 07:15:19 -0800
X-Trace: forums-1-dub 1076944519 172.31.142.236 (16 Feb 2004 07:15:19 -0800)
X-Original-Trace: 16 Feb 2004 07:15:19 -0800, rwaywell-pc.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2513
Article PK: 5651

So what you have observed and described is a situation in which there is no
deadlock in the database engine as confirmed by the sa_conn_info output,
however there does appear to be a deadlock in the application.

--
-----------------------------------------------
Robert Waywell
Sybase Adaptive Server Anywhere Developer - Version 8
Sybase Certified Professional

Sybase's iAnywhere Solutions

Please respond ONLY to newsgroup

EBF's and Patches: http://downloads.sybase.com
choose SQL Anywhere Studio >> change 'time frame' to all

To Submit Bug Reports:
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
ug

SQL Anywhere Studio Supported Platforms and Support Status
http://my.sybase.com/detail?id=1002288

Whitepapers, TechDocs, and bug fixes are all available through the iAnywhere
Developer Community at www.ianywhere.com/developer

<modan> wrote in message news:4030b328.234a.846930886@sybase.com...
> Just to be clear, our program invokes 2 threads. The same
> program can be run in a different mode where the program
> only runs one of the threads, allowing us to run two
> instances. When we run in this mode (two applications
> rather than two threads) then the deadlock detection works,
> and behaves as expected.
>
> > What settings do you use for the BLOCKING and
> > BLOCKING_TIMEOUT options?
> >
> > Is it possible that the difference between 1250 and 1252
> > is that a different connection is being chosen as the
> > "victim" when a deadlock is detected?
> >
> > Breck
> >
> > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> >
> > >Hi, I'm going to be taking this up, as Dave is off on
> > >holiday.
> > >
> > >Anyway, I have confirmed that the build is 1252.
> > >
> > >In answer to your other question, sa_conn_info shows one
> > >blocked connection, not two.
> > >
> > >Both show ReqType of STMT_EXECUTE, and identical
> > >lastreqtime, and UncmtOps = 0. The "unblocked" one shows
> > a >LastIdle greater than the blocked one by either 1 or 2.
> > >
> > >Modan
> > >
> > >> So sa_conn_info showed two blocked connections, blocked
> > on >> each other? That's a situation that should never
> > actually >> be viewable IMHO.
> > >>
> > >> Are you sure you're on 1252, not build 1270?
> > >>
> > >> Please excuse my dumb questions, but I am chasing a
> > >> related issue.
> > >>
> > >> Breck
> > >>
> > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > >>
> > >> >I found the "Deadlock Detected" worked when the
> > threads >> were >in separate applications but not when in
> > the same >> >application
> > >> >
> > >> >It did show a lock Cycle, Only 3 out of the 4 Updates
> > to >> the >database were processed by the server.
> > >> >
> > >> >Dave
> > >> >
> > >> >
> > >> >> FWIW I just ran a deadlock test using two
> > connections >> and >> got "Deadlock detected" as expected,
> > in build 1252. >> >>
> > >> >> What does sa_conn_info show when you run it on a 3rd
> > >> >> connection? Does it show an actual lock cycle
> > >> (deadlock)? >>
> > >> >> Breck
> > >> >>
> > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> > >> >>
> > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > Build >> 1108 >> >
> > >> >> >When Two Threads created a DeadLock, I get the
> > Correct >> >> >Deadlock Error returned for one of the
> > Threads, >> >> >
> > >> >> >However when I install the Express Bug Fix, Update
> > of >> >> 9.0.0 >to Build 1252
> > >> >> >
> > >> >> >I no longer get the DeadLock Error, The ASA
> > Commands >> just >> >Hang.
> > >> >> >
> > >> >> >Has anyone else seen this ?
> > >> >> >
> > >> >> >Thanks in advance
> > >> >> >
> > >> >> >Dave
> > >> >>
> > >> >> --
> > >> >> bcarter@risingroad.com
> > >> >> Mobile and Distributed Enterprise Database
> > Applications >> >> www.risingroad.com
> > >>
> > >> --
> > >> bcarter@risingroad.com
> > >> Mobile and Distributed Enterprise Database Applications
> > >> www.risingroad.com
> >
> > --
> > bcarter@risingroad.com
> > Mobile and Distributed Enterprise Database Applications
> > www.risingroad.com


modan Posted on 2004-02-16 15:29:24.0Z
Sender: 1d6f.4030e39d.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <4030e475.1d83.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub>
X-Original-NNTP-Posting-Host: 10.22.241.42
X-Original-Trace: 16 Feb 2004 07:40:37 -0800, 10.22.241.42
Lines: 141
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 16 Feb 2004 07:10:57 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 16 Feb 2004 07:29:24 -0800
X-Trace: forums-1-dub 1076945364 10.22.108.75 (16 Feb 2004 07:29:24 -0800)
X-Original-Trace: 16 Feb 2004 07:29:24 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2514
Article PK: 5652

Exactly!

I have just added some extra tracing in the application, and
confirmed that both threads execute the line

cmd2.ExecuteNonQuery();

So the problem might be related to the .Net Data Provider
rather than the Database Engine. Does this get changed as
part of an EBF?

> So what you have observed and described is a situation in
> which there is no deadlock in the database engine as
> confirmed by the sa_conn_info output, however there does
> appear to be a deadlock in the application.
>
> --
> -----------------------------------------------
> Robert Waywell
> Sybase Adaptive Server Anywhere Developer - Version 8
> Sybase Certified Professional
>
> Sybase's iAnywhere Solutions
>
> Please respond ONLY to newsgroup
>
> EBF's and Patches: http://downloads.sybase.com
> choose SQL Anywhere Studio >> change 'time frame' to all
>
> To Submit Bug Reports:
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> ug
>
> SQL Anywhere Studio Supported Platforms and Support Status
> http://my.sybase.com/detail?id=1002288
>
> Whitepapers, TechDocs, and bug fixes are all available
> through the iAnywhere Developer Community at
> www.ianywhere.com/developer
>
> <modan> wrote in message
> > news:4030b328.234a.846930886@sybase.com... Just to be
> > clear, our program invokes 2 threads. The same program
> > can be run in a different mode where the program only
> > runs one of the threads, allowing us to run two
> > instances. When we run in this mode (two applications
> rather than two threads) then the deadlock detection works
> > , and behaves as expected.
> >
> > > What settings do you use for the BLOCKING and
> > > BLOCKING_TIMEOUT options?
> > >
> > > Is it possible that the difference between 1250 and
> > > 1252 is that a different connection is being chosen as
> > > the "victim" when a deadlock is detected?
> > >
> > > Breck
> > >
> > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > >
> > > >Hi, I'm going to be taking this up, as Dave is off on
> > > >holiday.
> > > >
> > > >Anyway, I have confirmed that the build is 1252.
> > > >
> > > >In answer to your other question, sa_conn_info shows
> > > one >blocked connection, not two.
> > > >
> > > >Both show ReqType of STMT_EXECUTE, and identical
> > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > shows a >LastIdle greater than the blocked one by
> > > either 1 or 2. >
> > > >Modan
> > > >
> > > >> So sa_conn_info showed two blocked connections,
> > > blocked on >> each other? That's a situation that
> > > should never actually >> be viewable IMHO.
> > > >>
> > > >> Are you sure you're on 1252, not build 1270?
> > > >>
> > > >> Please excuse my dumb questions, but I am chasing a
> > > >> related issue.
> > > >>
> > > >> Breck
> > > >>
> > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > >>
> > > >> >I found the "Deadlock Detected" worked when the
> > > threads >> were >in separate applications but not when
> > > in the same >> >application
> > > >> >
> > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > Updates to >> the >database were processed by the
> > > server. >> >
> > > >> >Dave
> > > >> >
> > > >> >
> > > >> >> FWIW I just ran a deadlock test using two
> > > connections >> and >> got "Deadlock detected" as
> > > expected, in build 1252. >> >>
> > > >> >> What does sa_conn_info show when you run it on a
> > > 3rd >> >> connection? Does it show an actual lock
> > > cycle >> (deadlock)? >>
> > > >> >> Breck
> > > >> >>
> > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > wrote: >> >>
> > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > Build >> 1108 >> >
> > > >> >> >When Two Threads created a DeadLock, I get the
> > > Correct >> >> >Deadlock Error returned for one of the
> > > Threads, >> >> >
> > > >> >> >However when I install the Express Bug Fix,
> > > Update of >> >> 9.0.0 >to Build 1252
> > > >> >> >
> > > >> >> >I no longer get the DeadLock Error, The ASA
> > > Commands >> just >> >Hang.
> > > >> >> >
> > > >> >> >Has anyone else seen this ?
> > > >> >> >
> > > >> >> >Thanks in advance
> > > >> >> >
> > > >> >> >Dave
> > > >> >>
> > > >> >> --
> > > >> >> bcarter@risingroad.com
> > > >> >> Mobile and Distributed Enterprise Database
> > > Applications >> >> www.risingroad.com
> > > >>
> > > >> --
> > > >> bcarter@risingroad.com
> > > >> Mobile and Distributed Enterprise Database
> > > Applications >> www.risingroad.com
> > >
> > > --
> > > bcarter@risingroad.com
> > > Mobile and Distributed Enterprise Database
> > > Applications www.risingroad.com
>
>


Robert Waywell Posted on 2004-02-17 14:08:33.0Z
From: "Robert Waywell" <nospam_rwaywell@ianywhere.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 174
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
X-Original-NNTP-Posting-Host: rwaywell-pc.sybase.com
Message-ID: <40322308@forums-2-dub>
X-Original-Trace: 17 Feb 2004 06:19:52 -0800, rwaywell-pc.sybase.com
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 17 Feb 2004 05:50:01 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 17 Feb 2004 06:08:33 -0800
X-Trace: forums-1-dub 1077026913 10.22.108.75 (17 Feb 2004 06:08:33 -0800)
X-Original-Trace: 17 Feb 2004 06:08:33 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2528
Article PK: 5666

You can check the build number on the dboledb9.dll to see if was one of the
files updated in the EBF. Personally I suspect that the issue is in the
application rather than in the provider. One possibility would be the
failure to handle the deadlock detected error on thread X and tell thread Y
to continue.

--
-----------------------------------------------
Robert Waywell
Sybase Adaptive Server Anywhere Developer - Version 8
Sybase Certified Professional

Sybase's iAnywhere Solutions

Please respond ONLY to newsgroup

EBF's and Patches: http://downloads.sybase.com
choose SQL Anywhere Studio >> change 'time frame' to all

To Submit Bug Reports:
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
ug

SQL Anywhere Studio Supported Platforms and Support Status
http://my.sybase.com/detail?id=1002288

Whitepapers, TechDocs, and bug fixes are all available through the iAnywhere
Developer Community at www.ianywhere.com/developer

<modan> wrote in message news:4030e475.1d83.846930886@sybase.com...
> Exactly!
>
> I have just added some extra tracing in the application, and
> confirmed that both threads execute the line
>
> cmd2.ExecuteNonQuery();
>
> So the problem might be related to the .Net Data Provider
> rather than the Database Engine. Does this get changed as
> part of an EBF?
>
> > So what you have observed and described is a situation in
> > which there is no deadlock in the database engine as
> > confirmed by the sa_conn_info output, however there does
> > appear to be a deadlock in the application.
> >
> > --
> > -----------------------------------------------
> > Robert Waywell
> > Sybase Adaptive Server Anywhere Developer - Version 8
> > Sybase Certified Professional
> >
> > Sybase's iAnywhere Solutions
> >
> > Please respond ONLY to newsgroup
> >
> > EBF's and Patches: http://downloads.sybase.com
> > choose SQL Anywhere Studio >> change 'time frame' to all
> >
> > To Submit Bug Reports:
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > ug
> >
> > SQL Anywhere Studio Supported Platforms and Support Status
> > http://my.sybase.com/detail?id=1002288
> >
> > Whitepapers, TechDocs, and bug fixes are all available
> > through the iAnywhere Developer Community at
> > www.ianywhere.com/developer
> >
> > <modan> wrote in message
> > > news:4030b328.234a.846930886@sybase.com... Just to be
> > > clear, our program invokes 2 threads. The same program
> > > can be run in a different mode where the program only
> > > runs one of the threads, allowing us to run two
> > > instances. When we run in this mode (two applications
> > rather than two threads) then the deadlock detection works
> > > , and behaves as expected.
> > >
> > > > What settings do you use for the BLOCKING and
> > > > BLOCKING_TIMEOUT options?
> > > >
> > > > Is it possible that the difference between 1250 and
> > > > 1252 is that a different connection is being chosen as
> > > > the "victim" when a deadlock is detected?
> > > >
> > > > Breck
> > > >
> > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > >
> > > > >Hi, I'm going to be taking this up, as Dave is off on
> > > > >holiday.
> > > > >
> > > > >Anyway, I have confirmed that the build is 1252.
> > > > >
> > > > >In answer to your other question, sa_conn_info shows
> > > > one >blocked connection, not two.
> > > > >
> > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > > shows a >LastIdle greater than the blocked one by
> > > > either 1 or 2. >
> > > > >Modan
> > > > >
> > > > >> So sa_conn_info showed two blocked connections,
> > > > blocked on >> each other? That's a situation that
> > > > should never actually >> be viewable IMHO.
> > > > >>
> > > > >> Are you sure you're on 1252, not build 1270?
> > > > >>
> > > > >> Please excuse my dumb questions, but I am chasing a
> > > > >> related issue.
> > > > >>
> > > > >> Breck
> > > > >>
> > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > > >>
> > > > >> >I found the "Deadlock Detected" worked when the
> > > > threads >> were >in separate applications but not when
> > > > in the same >> >application
> > > > >> >
> > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > Updates to >> the >database were processed by the
> > > > server. >> >
> > > > >> >Dave
> > > > >> >
> > > > >> >
> > > > >> >> FWIW I just ran a deadlock test using two
> > > > connections >> and >> got "Deadlock detected" as
> > > > expected, in build 1252. >> >>
> > > > >> >> What does sa_conn_info show when you run it on a
> > > > 3rd >> >> connection? Does it show an actual lock
> > > > cycle >> (deadlock)? >>
> > > > >> >> Breck
> > > > >> >>
> > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > wrote: >> >>
> > > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > > Build >> 1108 >> >
> > > > >> >> >When Two Threads created a DeadLock, I get the
> > > > Correct >> >> >Deadlock Error returned for one of the
> > > > Threads, >> >> >
> > > > >> >> >However when I install the Express Bug Fix,
> > > > Update of >> >> 9.0.0 >to Build 1252
> > > > >> >> >
> > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > Commands >> just >> >Hang.
> > > > >> >> >
> > > > >> >> >Has anyone else seen this ?
> > > > >> >> >
> > > > >> >> >Thanks in advance
> > > > >> >> >
> > > > >> >> >Dave
> > > > >> >>
> > > > >> >> --
> > > > >> >> bcarter@risingroad.com
> > > > >> >> Mobile and Distributed Enterprise Database
> > > > Applications >> >> www.risingroad.com
> > > > >>
> > > > >> --
> > > > >> bcarter@risingroad.com
> > > > >> Mobile and Distributed Enterprise Database
> > > > Applications >> www.risingroad.com
> > > >
> > > > --
> > > > bcarter@risingroad.com
> > > > Mobile and Distributed Enterprise Database
> > > > Applications www.risingroad.com
> >
> >


modan Posted on 2004-02-19 10:48:50.0Z
Sender: 5261.403493ab.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <40349492.5267.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com><40322308@forums-2-dub>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 19 Feb 2004 02:48:50 -0800
X-Trace: forums-1-dub 1077187730 10.22.241.41 (19 Feb 2004 02:48:50 -0800)
X-Original-Trace: 19 Feb 2004 02:48:50 -0800, 10.22.241.41
Lines: 191
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2547
Article PK: 5685

Actually we are using the Native provider
(iAnywhere.Data.AsaClient.dll).

I have just run more experiments on various versions, and
have found the following.

Works on 9.0.0.1108 (of iAnywhere.Data.AsaClient.dll)

Does not work on versions up to and including 9.0.1.1751
(the current)

> You can check the build number on the dboledb9.dll to see
> if was one of the files updated in the EBF. Personally I
> suspect that the issue is in the application rather than
> in the provider. One possibility would be the failure to
> handle the deadlock detected error on thread X and tell
> thread Y to continue.
>
> --
> -----------------------------------------------
> Robert Waywell
> Sybase Adaptive Server Anywhere Developer - Version 8
> Sybase Certified Professional
>
> Sybase's iAnywhere Solutions
>
> Please respond ONLY to newsgroup
>
> EBF's and Patches: http://downloads.sybase.com
> choose SQL Anywhere Studio >> change 'time frame' to all
>
> To Submit Bug Reports:
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> ug
>
> SQL Anywhere Studio Supported Platforms and Support Status
> http://my.sybase.com/detail?id=1002288
>
> Whitepapers, TechDocs, and bug fixes are all available
> through the iAnywhere Developer Community at
> www.ianywhere.com/developer
>
> <modan> wrote in message
> > news:4030e475.1d83.846930886@sybase.com... Exactly!
> >
> > I have just added some extra tracing in the application,
> > and confirmed that both threads execute the line
> >
> > cmd2.ExecuteNonQuery();
> >
> > So the problem might be related to the .Net Data
> > Provider rather than the Database Engine. Does this get
> > changed as part of an EBF?
> >
> > > So what you have observed and described is a situation
> > > in which there is no deadlock in the database engine
> > > as confirmed by the sa_conn_info output, however there
> > > does appear to be a deadlock in the application.
> > >
> > > --
> > > -----------------------------------------------
> > > Robert Waywell
> > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > Sybase Certified Professional
> > >
> > > Sybase's iAnywhere Solutions
> > >
> > > Please respond ONLY to newsgroup
> > >
> > > EBF's and Patches: http://downloads.sybase.com
> > > choose SQL Anywhere Studio >> change 'time frame' to
> > all >
> > > To Submit Bug Reports:
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > ug
> > >
> > > SQL Anywhere Studio Supported Platforms and Support
> > > Status http://my.sybase.com/detail?id=1002288
> > >
> > > Whitepapers, TechDocs, and bug fixes are all available
> > > through the iAnywhere Developer Community at
> > > www.ianywhere.com/developer
> > >
> > > <modan> wrote in message
> > > > news:4030b328.234a.846930886@sybase.com... Just to
> > > > be clear, our program invokes 2 threads. The same
> > > > program can be run in a different mode where the
> > > > program only runs one of the threads, allowing us to
> > > > run two instances. When we run in this mode (two
> > > applications rather than two threads) then the
> > > > deadlock detection works , and behaves as expected.
> > > >
> > > > > What settings do you use for the BLOCKING and
> > > > > BLOCKING_TIMEOUT options?
> > > > >
> > > > > Is it possible that the difference between 1250
> > > > > and 1252 is that a different connection is being
> > > > > chosen as the "victim" when a deadlock is
> > > > detected? >
> > > > > Breck
> > > > >
> > > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > > >
> > > > > >Hi, I'm going to be taking this up, as Dave is
> > > > > off on >holiday.
> > > > > >
> > > > > >Anyway, I have confirmed that the build is 1252.
> > > > > >
> > > > > >In answer to your other question, sa_conn_info
> > > > > shows one >blocked connection, not two.
> > > > > >
> > > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > > >lastreqtime, and UncmtOps = 0. The "unblocked"
> > > > > one shows a >LastIdle greater than the blocked one
> > > > > by either 1 or 2. >
> > > > > >Modan
> > > > > >
> > > > > >> So sa_conn_info showed two blocked connections,
> > > > > blocked on >> each other? That's a situation that
> > > > > should never actually >> be viewable IMHO.
> > > > > >>
> > > > > >> Are you sure you're on 1252, not build 1270?
> > > > > >>
> > > > > >> Please excuse my dumb questions, but I am
> > > > > chasing a >> related issue.
> > > > > >>
> > > > > >> Breck
> > > > > >>
> > > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage
> > > > > wrote: >>
> > > > > >> >I found the "Deadlock Detected" worked when
> > > > > the threads >> were >in separate applications but
> > > > > not when in the same >> >application
> > > > > >> >
> > > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > > Updates to >> the >database were processed by the
> > > > > server. >> >
> > > > > >> >Dave
> > > > > >> >
> > > > > >> >
> > > > > >> >> FWIW I just ran a deadlock test using two
> > > > > connections >> and >> got "Deadlock detected" as
> > > > > expected, in build 1252. >> >>
> > > > > >> >> What does sa_conn_info show when you run it
> > > > > on a 3rd >> >> connection? Does it show an actual
> > > > > lock cycle >> (deadlock)? >>
> > > > > >> >> Breck
> > > > > >> >>
> > > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > > wrote: >> >>
> > > > > >> >> >Working with Sybase SQL Anywhere Verion
> > > > > 9.0.0. Build >> 1108 >> >
> > > > > >> >> >When Two Threads created a DeadLock, I get
> > > > > the Correct >> >> >Deadlock Error returned for one
> > > > > of the Threads, >> >> >
> > > > > >> >> >However when I install the Express Bug Fix,
> > > > > Update of >> >> 9.0.0 >to Build 1252
> > > > > >> >> >
> > > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > > Commands >> just >> >Hang.
> > > > > >> >> >
> > > > > >> >> >Has anyone else seen this ?
> > > > > >> >> >
> > > > > >> >> >Thanks in advance
> > > > > >> >> >
> > > > > >> >> >Dave
> > > > > >> >>
> > > > > >> >> --
> > > > > >> >> bcarter@risingroad.com
> > > > > >> >> Mobile and Distributed Enterprise Database
> > > > > Applications >> >> www.risingroad.com
> > > > > >>
> > > > > >> --
> > > > > >> bcarter@risingroad.com
> > > > > >> Mobile and Distributed Enterprise Database
> > > > > Applications >> www.risingroad.com
> > > > >
> > > > > --
> > > > > bcarter@risingroad.com
> > > > > Mobile and Distributed Enterprise Database
> > > > > Applications www.risingroad.com
> > >
> > >
>
>


modan Posted on 2004-02-24 16:53:56.0Z
From: "modan" <me@here.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 206
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1158
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
X-Original-NNTP-Posting-Host: tap.theautomationpartnership.co.uk
Message-ID: <403b81b1@forums-2-dub>
X-Original-Trace: 24 Feb 2004 08:54:09 -0800, tap.theautomationpartnership.co.uk
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 24 Feb 2004 08:53:24 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 24 Feb 2004 08:53:56 -0800
X-Trace: forums-1-dub 1077641636 10.22.108.75 (24 Feb 2004 08:53:56 -0800)
X-Original-Trace: 24 Feb 2004 08:53:56 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2618
Article PK: 6101

Things have gone a bit quiet on this thread. I guess that means no one has
any ideas?

FWIW this is what I think is happening.

Thread 1 updates Table A
Thread 1 gets interupted by Thread 2
Thread 2 updates Table B
Thread 2 tries to update Table A, but blocks on Thread 1's update ...and
this is where things seem to go wrong.
What should happen is that Thread 2 relinquishes control back to Thread 1,
which then tries to update Table B, and the engine detects a deadlock
However, we know that this is not happening from the logging that we have
been collecting, so presumably this means that something has happened within
the internals of the .Net Provider (iAnywhere) that is preventing the Thread
1 from executing in later versions.

Would it be useful for me to provide the source code that exhibits this
behaviour? I think it is clear that this is not a problem with our
application since it worked with an earlier Sybase build, but perhaps it
would help to make it clearer where things are going wrong

"Robert Waywell" <nospam_rwaywell@ianywhere.com> wrote in message
news:40322308@forums-2-dub...
> You can check the build number on the dboledb9.dll to see if was one of
the
> files updated in the EBF. Personally I suspect that the issue is in the
> application rather than in the provider. One possibility would be the
> failure to handle the deadlock detected error on thread X and tell thread
Y
> to continue.
>
> --
> -----------------------------------------------
> Robert Waywell
> Sybase Adaptive Server Anywhere Developer - Version 8
> Sybase Certified Professional
>
> Sybase's iAnywhere Solutions
>
> Please respond ONLY to newsgroup
>
> EBF's and Patches: http://downloads.sybase.com
> choose SQL Anywhere Studio >> change 'time frame' to all
>
> To Submit Bug Reports:
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> ug
>
> SQL Anywhere Studio Supported Platforms and Support Status
> http://my.sybase.com/detail?id=1002288
>
> Whitepapers, TechDocs, and bug fixes are all available through the
iAnywhere
> Developer Community at www.ianywhere.com/developer
>
> <modan> wrote in message news:4030e475.1d83.846930886@sybase.com...
> > Exactly!
> >
> > I have just added some extra tracing in the application, and
> > confirmed that both threads execute the line
> >
> > cmd2.ExecuteNonQuery();
> >
> > So the problem might be related to the .Net Data Provider
> > rather than the Database Engine. Does this get changed as
> > part of an EBF?
> >
> > > So what you have observed and described is a situation in
> > > which there is no deadlock in the database engine as
> > > confirmed by the sa_conn_info output, however there does
> > > appear to be a deadlock in the application.
> > >
> > > --
> > > -----------------------------------------------
> > > Robert Waywell
> > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > Sybase Certified Professional
> > >
> > > Sybase's iAnywhere Solutions
> > >
> > > Please respond ONLY to newsgroup
> > >
> > > EBF's and Patches: http://downloads.sybase.com
> > > choose SQL Anywhere Studio >> change 'time frame' to all
> > >
> > > To Submit Bug Reports:
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > ug
> > >
> > > SQL Anywhere Studio Supported Platforms and Support Status
> > > http://my.sybase.com/detail?id=1002288
> > >
> > > Whitepapers, TechDocs, and bug fixes are all available
> > > through the iAnywhere Developer Community at
> > > www.ianywhere.com/developer
> > >
> > > <modan> wrote in message
> > > > news:4030b328.234a.846930886@sybase.com... Just to be
> > > > clear, our program invokes 2 threads. The same program
> > > > can be run in a different mode where the program only
> > > > runs one of the threads, allowing us to run two
> > > > instances. When we run in this mode (two applications
> > > rather than two threads) then the deadlock detection works
> > > > , and behaves as expected.
> > > >
> > > > > What settings do you use for the BLOCKING and
> > > > > BLOCKING_TIMEOUT options?
> > > > >
> > > > > Is it possible that the difference between 1250 and
> > > > > 1252 is that a different connection is being chosen as
> > > > > the "victim" when a deadlock is detected?
> > > > >
> > > > > Breck
> > > > >
> > > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > > >
> > > > > >Hi, I'm going to be taking this up, as Dave is off on
> > > > > >holiday.
> > > > > >
> > > > > >Anyway, I have confirmed that the build is 1252.
> > > > > >
> > > > > >In answer to your other question, sa_conn_info shows
> > > > > one >blocked connection, not two.
> > > > > >
> > > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > > > shows a >LastIdle greater than the blocked one by
> > > > > either 1 or 2. >
> > > > > >Modan
> > > > > >
> > > > > >> So sa_conn_info showed two blocked connections,
> > > > > blocked on >> each other? That's a situation that
> > > > > should never actually >> be viewable IMHO.
> > > > > >>
> > > > > >> Are you sure you're on 1252, not build 1270?
> > > > > >>
> > > > > >> Please excuse my dumb questions, but I am chasing a
> > > > > >> related issue.
> > > > > >>
> > > > > >> Breck
> > > > > >>
> > > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > > > >>
> > > > > >> >I found the "Deadlock Detected" worked when the
> > > > > threads >> were >in separate applications but not when
> > > > > in the same >> >application
> > > > > >> >
> > > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > > Updates to >> the >database were processed by the
> > > > > server. >> >
> > > > > >> >Dave
> > > > > >> >
> > > > > >> >
> > > > > >> >> FWIW I just ran a deadlock test using two
> > > > > connections >> and >> got "Deadlock detected" as
> > > > > expected, in build 1252. >> >>
> > > > > >> >> What does sa_conn_info show when you run it on a
> > > > > 3rd >> >> connection? Does it show an actual lock
> > > > > cycle >> (deadlock)? >>
> > > > > >> >> Breck
> > > > > >> >>
> > > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > > wrote: >> >>
> > > > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > > > Build >> 1108 >> >
> > > > > >> >> >When Two Threads created a DeadLock, I get the
> > > > > Correct >> >> >Deadlock Error returned for one of the
> > > > > Threads, >> >> >
> > > > > >> >> >However when I install the Express Bug Fix,
> > > > > Update of >> >> 9.0.0 >to Build 1252
> > > > > >> >> >
> > > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > > Commands >> just >> >Hang.
> > > > > >> >> >
> > > > > >> >> >Has anyone else seen this ?
> > > > > >> >> >
> > > > > >> >> >Thanks in advance
> > > > > >> >> >
> > > > > >> >> >Dave
> > > > > >> >>
> > > > > >> >> --
> > > > > >> >> bcarter@risingroad.com
> > > > > >> >> Mobile and Distributed Enterprise Database
> > > > > Applications >> >> www.risingroad.com
> > > > > >>
> > > > > >> --
> > > > > >> bcarter@risingroad.com
> > > > > >> Mobile and Distributed Enterprise Database
> > > > > Applications >> www.risingroad.com
> > > > >
> > > > > --
> > > > > bcarter@risingroad.com
> > > > > Mobile and Distributed Enterprise Database
> > > > > Applications www.risingroad.com
> > >
> > >
>
>


Chris Keating (iAnywhere Solutions) Posted on 2004-02-24 17:04:28.0Z
From: "Chris Keating \(iAnywhere Solutions\)" <FightSpam_keating@iAnywhere.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub> <403b81b1@forums-2-dub>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 257
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1158
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
NNTP-Posting-Host: keating-xp.sybase.com
X-Original-NNTP-Posting-Host: keating-xp.sybase.com
Message-ID: <403b841c$1@forums-1-dub>
Date: 24 Feb 2004 09:04:28 -0800
X-Trace: forums-1-dub 1077642268 172.31.141.1 (24 Feb 2004 09:04:28 -0800)
X-Original-Trace: 24 Feb 2004 09:04:28 -0800, keating-xp.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2619
Article PK: 6092

You'll probably need to contact Technical Support to get this issue
resolved. Within North America, you can reach Technical Support by calling
1-800-8SYBASE. If you are located outside of North America, you can find
the contact information for your local Sybase Support Center by going to the
Sybase home page at http://www.sybase.com, clicking on the "Support" link at
the top of the page, and then choosing "Support Centers" from the left hand
frame.

You can always report a bug for free by visiting CaseXpress at
http://casexpress.sybase.com/cx/cx.stm. When submitting calls through
CaseXpress, please make sure to give an accurate description of the steps
you take to reproduce the bug, and include any files that may be needed
(database, source code) to reproduce the problem.


--

Chris Keating
Sybase Adaptive Server Anywhere Professional Version 8

****************************************************************************
*
Sign up today for your copy of the SQL Anywhere Studio 9 Developer Edition
and try out the market-leading database for mobile, embedded and small to
medium sized business environments for free!

http://www.ianywhere.com/promos/deved/index.html

****************************************************************************
*

iAnywhere Solutions http://www.iAnywhere.com

** Please only post to the newsgroup

** Whitepapers can be found at http://www.iAnywhere.com/developer
** EBFs can be found at http://downloads.sybase.com/swx/sdmain.stm
** Use CaseXpress to report bugs http://casexpress.sybase.com

****************************************************************************
*

"modan" <me@here.com> wrote in message news:403b81b1@forums-2-dub...
> Things have gone a bit quiet on this thread. I guess that means no one
has
> any ideas?
>
> FWIW this is what I think is happening.
>
> Thread 1 updates Table A
> Thread 1 gets interupted by Thread 2
> Thread 2 updates Table B
> Thread 2 tries to update Table A, but blocks on Thread 1's update ...and
> this is where things seem to go wrong.
> What should happen is that Thread 2 relinquishes control back to Thread 1,
> which then tries to update Table B, and the engine detects a deadlock
> However, we know that this is not happening from the logging that we have
> been collecting, so presumably this means that something has happened
within
> the internals of the .Net Provider (iAnywhere) that is preventing the
Thread
> 1 from executing in later versions.
>
> Would it be useful for me to provide the source code that exhibits this
> behaviour? I think it is clear that this is not a problem with our
> application since it worked with an earlier Sybase build, but perhaps it
> would help to make it clearer where things are going wrong
>
>
> "Robert Waywell" <nospam_rwaywell@ianywhere.com> wrote in message
> news:40322308@forums-2-dub...
> > You can check the build number on the dboledb9.dll to see if was one of
> the
> > files updated in the EBF. Personally I suspect that the issue is in the
> > application rather than in the provider. One possibility would be the
> > failure to handle the deadlock detected error on thread X and tell
thread
> Y
> > to continue.
> >
> > --
> > -----------------------------------------------
> > Robert Waywell
> > Sybase Adaptive Server Anywhere Developer - Version 8
> > Sybase Certified Professional
> >
> > Sybase's iAnywhere Solutions
> >
> > Please respond ONLY to newsgroup
> >
> > EBF's and Patches: http://downloads.sybase.com
> > choose SQL Anywhere Studio >> change 'time frame' to all
> >
> > To Submit Bug Reports:
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > ug
> >
> > SQL Anywhere Studio Supported Platforms and Support Status
> > http://my.sybase.com/detail?id=1002288
> >
> > Whitepapers, TechDocs, and bug fixes are all available through the
> iAnywhere
> > Developer Community at www.ianywhere.com/developer
> >
> > <modan> wrote in message news:4030e475.1d83.846930886@sybase.com...
> > > Exactly!
> > >
> > > I have just added some extra tracing in the application, and
> > > confirmed that both threads execute the line
> > >
> > > cmd2.ExecuteNonQuery();
> > >
> > > So the problem might be related to the .Net Data Provider
> > > rather than the Database Engine. Does this get changed as
> > > part of an EBF?
> > >
> > > > So what you have observed and described is a situation in
> > > > which there is no deadlock in the database engine as
> > > > confirmed by the sa_conn_info output, however there does
> > > > appear to be a deadlock in the application.
> > > >
> > > > --
> > > > -----------------------------------------------
> > > > Robert Waywell
> > > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > > Sybase Certified Professional
> > > >
> > > > Sybase's iAnywhere Solutions
> > > >
> > > > Please respond ONLY to newsgroup
> > > >
> > > > EBF's and Patches: http://downloads.sybase.com
> > > > choose SQL Anywhere Studio >> change 'time frame' to all
> > > >
> > > > To Submit Bug Reports:
> > > >
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > > ug
> > > >
> > > > SQL Anywhere Studio Supported Platforms and Support Status
> > > > http://my.sybase.com/detail?id=1002288
> > > >
> > > > Whitepapers, TechDocs, and bug fixes are all available
> > > > through the iAnywhere Developer Community at
> > > > www.ianywhere.com/developer
> > > >
> > > > <modan> wrote in message
> > > > > news:4030b328.234a.846930886@sybase.com... Just to be
> > > > > clear, our program invokes 2 threads. The same program
> > > > > can be run in a different mode where the program only
> > > > > runs one of the threads, allowing us to run two
> > > > > instances. When we run in this mode (two applications
> > > > rather than two threads) then the deadlock detection works
> > > > > , and behaves as expected.
> > > > >
> > > > > > What settings do you use for the BLOCKING and
> > > > > > BLOCKING_TIMEOUT options?
> > > > > >
> > > > > > Is it possible that the difference between 1250 and
> > > > > > 1252 is that a different connection is being chosen as
> > > > > > the "victim" when a deadlock is detected?
> > > > > >
> > > > > > Breck
> > > > > >
> > > > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > > > >
> > > > > > >Hi, I'm going to be taking this up, as Dave is off on
> > > > > > >holiday.
> > > > > > >
> > > > > > >Anyway, I have confirmed that the build is 1252.
> > > > > > >
> > > > > > >In answer to your other question, sa_conn_info shows
> > > > > > one >blocked connection, not two.
> > > > > > >
> > > > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > > > > shows a >LastIdle greater than the blocked one by
> > > > > > either 1 or 2. >
> > > > > > >Modan
> > > > > > >
> > > > > > >> So sa_conn_info showed two blocked connections,
> > > > > > blocked on >> each other? That's a situation that
> > > > > > should never actually >> be viewable IMHO.
> > > > > > >>
> > > > > > >> Are you sure you're on 1252, not build 1270?
> > > > > > >>
> > > > > > >> Please excuse my dumb questions, but I am chasing a
> > > > > > >> related issue.
> > > > > > >>
> > > > > > >> Breck
> > > > > > >>
> > > > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > > > > >>
> > > > > > >> >I found the "Deadlock Detected" worked when the
> > > > > > threads >> were >in separate applications but not when
> > > > > > in the same >> >application
> > > > > > >> >
> > > > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > > > Updates to >> the >database were processed by the
> > > > > > server. >> >
> > > > > > >> >Dave
> > > > > > >> >
> > > > > > >> >
> > > > > > >> >> FWIW I just ran a deadlock test using two
> > > > > > connections >> and >> got "Deadlock detected" as
> > > > > > expected, in build 1252. >> >>
> > > > > > >> >> What does sa_conn_info show when you run it on a
> > > > > > 3rd >> >> connection? Does it show an actual lock
> > > > > > cycle >> (deadlock)? >>
> > > > > > >> >> Breck
> > > > > > >> >>
> > > > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > > > wrote: >> >>
> > > > > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > > > > Build >> 1108 >> >
> > > > > > >> >> >When Two Threads created a DeadLock, I get the
> > > > > > Correct >> >> >Deadlock Error returned for one of the
> > > > > > Threads, >> >> >
> > > > > > >> >> >However when I install the Express Bug Fix,
> > > > > > Update of >> >> 9.0.0 >to Build 1252
> > > > > > >> >> >
> > > > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > > > Commands >> just >> >Hang.
> > > > > > >> >> >
> > > > > > >> >> >Has anyone else seen this ?
> > > > > > >> >> >
> > > > > > >> >> >Thanks in advance
> > > > > > >> >> >
> > > > > > >> >> >Dave
> > > > > > >> >>
> > > > > > >> >> --
> > > > > > >> >> bcarter@risingroad.com
> > > > > > >> >> Mobile and Distributed Enterprise Database
> > > > > > Applications >> >> www.risingroad.com
> > > > > > >>
> > > > > > >> --
> > > > > > >> bcarter@risingroad.com
> > > > > > >> Mobile and Distributed Enterprise Database
> > > > > > Applications >> www.risingroad.com
> > > > > >
> > > > > > --
> > > > > > bcarter@risingroad.com
> > > > > > Mobile and Distributed Enterprise Database
> > > > > > Applications www.risingroad.com
> > > >
> > > >
> >
> >
>
>


modan Posted on 2004-02-24 17:14:17.0Z
From: "modan" <me@here.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub> <403b81b1@forums-2-dub> <403b841c$1@forums-1-dub>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 276
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1158
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
X-Original-NNTP-Posting-Host: tap.theautomationpartnership.co.uk
Message-ID: <403b8676@forums-2-dub>
X-Original-Trace: 24 Feb 2004 09:14:30 -0800, tap.theautomationpartnership.co.uk
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 24 Feb 2004 09:13:45 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 24 Feb 2004 09:14:17 -0800
X-Trace: forums-1-dub 1077642857 10.22.108.75 (24 Feb 2004 09:14:17 -0800)
X-Original-Trace: 24 Feb 2004 09:14:17 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2620
Article PK: 6094

OK, thanks

"Chris Keating (iAnywhere Solutions)" <FightSpam_keating@iAnywhere.com>

wrote in message news:403b841c$1@forums-1-dub...
> You'll probably need to contact Technical Support to get this issue
> resolved. Within North America, you can reach Technical Support by
calling
> 1-800-8SYBASE. If you are located outside of North America, you can find
> the contact information for your local Sybase Support Center by going to
the
> Sybase home page at http://www.sybase.com, clicking on the "Support" link
at
> the top of the page, and then choosing "Support Centers" from the left
hand
> frame.
>
> You can always report a bug for free by visiting CaseXpress at
> http://casexpress.sybase.com/cx/cx.stm. When submitting calls through
> CaseXpress, please make sure to give an accurate description of the steps
> you take to reproduce the bug, and include any files that may be needed
> (database, source code) to reproduce the problem.
>
>
> --
>
> Chris Keating
> Sybase Adaptive Server Anywhere Professional Version 8
>
>
****************************************************************************
> *
> Sign up today for your copy of the SQL Anywhere Studio 9 Developer Edition
> and try out the market-leading database for mobile, embedded and small to
> medium sized business environments for free!
>
> http://www.ianywhere.com/promos/deved/index.html
>
>
****************************************************************************
> *
>
> iAnywhere Solutions http://www.iAnywhere.com
>
> ** Please only post to the newsgroup
>
> ** Whitepapers can be found at http://www.iAnywhere.com/developer
> ** EBFs can be found at http://downloads.sybase.com/swx/sdmain.stm
> ** Use CaseXpress to report bugs http://casexpress.sybase.com
>
>
****************************************************************************
> *
>
> "modan" <me@here.com> wrote in message news:403b81b1@forums-2-dub...
> > Things have gone a bit quiet on this thread. I guess that means no one
> has
> > any ideas?
> >
> > FWIW this is what I think is happening.
> >
> > Thread 1 updates Table A
> > Thread 1 gets interupted by Thread 2
> > Thread 2 updates Table B
> > Thread 2 tries to update Table A, but blocks on Thread 1's update ...and
> > this is where things seem to go wrong.
> > What should happen is that Thread 2 relinquishes control back to Thread
1,
> > which then tries to update Table B, and the engine detects a deadlock
> > However, we know that this is not happening from the logging that we
have
> > been collecting, so presumably this means that something has happened
> within
> > the internals of the .Net Provider (iAnywhere) that is preventing the
> Thread
> > 1 from executing in later versions.
> >
> > Would it be useful for me to provide the source code that exhibits this
> > behaviour? I think it is clear that this is not a problem with our
> > application since it worked with an earlier Sybase build, but perhaps it
> > would help to make it clearer where things are going wrong
> >
> >
> > "Robert Waywell" <nospam_rwaywell@ianywhere.com> wrote in message
> > news:40322308@forums-2-dub...
> > > You can check the build number on the dboledb9.dll to see if was one
of
> > the
> > > files updated in the EBF. Personally I suspect that the issue is in
the
> > > application rather than in the provider. One possibility would be the
> > > failure to handle the deadlock detected error on thread X and tell
> thread
> > Y
> > > to continue.
> > >
> > > --
> > > -----------------------------------------------
> > > Robert Waywell
> > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > Sybase Certified Professional
> > >
> > > Sybase's iAnywhere Solutions
> > >
> > > Please respond ONLY to newsgroup
> > >
> > > EBF's and Patches: http://downloads.sybase.com
> > > choose SQL Anywhere Studio >> change 'time frame' to all
> > >
> > > To Submit Bug Reports:
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > ug
> > >
> > > SQL Anywhere Studio Supported Platforms and Support Status
> > > http://my.sybase.com/detail?id=1002288
> > >
> > > Whitepapers, TechDocs, and bug fixes are all available through the
> > iAnywhere
> > > Developer Community at www.ianywhere.com/developer
> > >
> > > <modan> wrote in message news:4030e475.1d83.846930886@sybase.com...
> > > > Exactly!
> > > >
> > > > I have just added some extra tracing in the application, and
> > > > confirmed that both threads execute the line
> > > >
> > > > cmd2.ExecuteNonQuery();
> > > >
> > > > So the problem might be related to the .Net Data Provider
> > > > rather than the Database Engine. Does this get changed as
> > > > part of an EBF?
> > > >
> > > > > So what you have observed and described is a situation in
> > > > > which there is no deadlock in the database engine as
> > > > > confirmed by the sa_conn_info output, however there does
> > > > > appear to be a deadlock in the application.
> > > > >
> > > > > --
> > > > > -----------------------------------------------
> > > > > Robert Waywell
> > > > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > > > Sybase Certified Professional
> > > > >
> > > > > Sybase's iAnywhere Solutions
> > > > >
> > > > > Please respond ONLY to newsgroup
> > > > >
> > > > > EBF's and Patches: http://downloads.sybase.com
> > > > > choose SQL Anywhere Studio >> change 'time frame' to all
> > > > >
> > > > > To Submit Bug Reports:
> > > > >
> > > >
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > > > ug
> > > > >
> > > > > SQL Anywhere Studio Supported Platforms and Support Status
> > > > > http://my.sybase.com/detail?id=1002288
> > > > >
> > > > > Whitepapers, TechDocs, and bug fixes are all available
> > > > > through the iAnywhere Developer Community at
> > > > > www.ianywhere.com/developer
> > > > >
> > > > > <modan> wrote in message
> > > > > > news:4030b328.234a.846930886@sybase.com... Just to be
> > > > > > clear, our program invokes 2 threads. The same program
> > > > > > can be run in a different mode where the program only
> > > > > > runs one of the threads, allowing us to run two
> > > > > > instances. When we run in this mode (two applications
> > > > > rather than two threads) then the deadlock detection works
> > > > > > , and behaves as expected.
> > > > > >
> > > > > > > What settings do you use for the BLOCKING and
> > > > > > > BLOCKING_TIMEOUT options?
> > > > > > >
> > > > > > > Is it possible that the difference between 1250 and
> > > > > > > 1252 is that a different connection is being chosen as
> > > > > > > the "victim" when a deadlock is detected?
> > > > > > >
> > > > > > > Breck
> > > > > > >
> > > > > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > > > > >
> > > > > > > >Hi, I'm going to be taking this up, as Dave is off on
> > > > > > > >holiday.
> > > > > > > >
> > > > > > > >Anyway, I have confirmed that the build is 1252.
> > > > > > > >
> > > > > > > >In answer to your other question, sa_conn_info shows
> > > > > > > one >blocked connection, not two.
> > > > > > > >
> > > > > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > > > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > > > > > shows a >LastIdle greater than the blocked one by
> > > > > > > either 1 or 2. >
> > > > > > > >Modan
> > > > > > > >
> > > > > > > >> So sa_conn_info showed two blocked connections,
> > > > > > > blocked on >> each other? That's a situation that
> > > > > > > should never actually >> be viewable IMHO.
> > > > > > > >>
> > > > > > > >> Are you sure you're on 1252, not build 1270?
> > > > > > > >>
> > > > > > > >> Please excuse my dumb questions, but I am chasing a
> > > > > > > >> related issue.
> > > > > > > >>
> > > > > > > >> Breck
> > > > > > > >>
> > > > > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > > > > > >>
> > > > > > > >> >I found the "Deadlock Detected" worked when the
> > > > > > > threads >> were >in separate applications but not when
> > > > > > > in the same >> >application
> > > > > > > >> >
> > > > > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > > > > Updates to >> the >database were processed by the
> > > > > > > server. >> >
> > > > > > > >> >Dave
> > > > > > > >> >
> > > > > > > >> >
> > > > > > > >> >> FWIW I just ran a deadlock test using two
> > > > > > > connections >> and >> got "Deadlock detected" as
> > > > > > > expected, in build 1252. >> >>
> > > > > > > >> >> What does sa_conn_info show when you run it on a
> > > > > > > 3rd >> >> connection? Does it show an actual lock
> > > > > > > cycle >> (deadlock)? >>
> > > > > > > >> >> Breck
> > > > > > > >> >>
> > > > > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > > > > wrote: >> >>
> > > > > > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > > > > > Build >> 1108 >> >
> > > > > > > >> >> >When Two Threads created a DeadLock, I get the
> > > > > > > Correct >> >> >Deadlock Error returned for one of the
> > > > > > > Threads, >> >> >
> > > > > > > >> >> >However when I install the Express Bug Fix,
> > > > > > > Update of >> >> 9.0.0 >to Build 1252
> > > > > > > >> >> >
> > > > > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > > > > Commands >> just >> >Hang.
> > > > > > > >> >> >
> > > > > > > >> >> >Has anyone else seen this ?
> > > > > > > >> >> >
> > > > > > > >> >> >Thanks in advance
> > > > > > > >> >> >
> > > > > > > >> >> >Dave
> > > > > > > >> >>
> > > > > > > >> >> --
> > > > > > > >> >> bcarter@risingroad.com
> > > > > > > >> >> Mobile and Distributed Enterprise Database
> > > > > > > Applications >> >> www.risingroad.com
> > > > > > > >>
> > > > > > > >> --
> > > > > > > >> bcarter@risingroad.com
> > > > > > > >> Mobile and Distributed Enterprise Database
> > > > > > > Applications >> www.risingroad.com
> > > > > > >
> > > > > > > --
> > > > > > > bcarter@risingroad.com
> > > > > > > Mobile and Distributed Enterprise Database
> > > > > > > Applications www.risingroad.com
> > > > >
> > > > >
> > >
> > >
> >
> >
>
>


modan Posted on 2004-03-24 10:00:22.0Z
From: "modan" <me@here.com>
Newsgroups: ianywhere.public.general
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com> <4030b328.234a.846930886@sybase.com><4030de87$1@forums-1-dub> <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub> <403b81b1@forums-2-dub> <403b841c$1@forums-1-dub> <403b8676@forums-2-dub>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 302
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1158
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
X-Original-NNTP-Posting-Host: tap.theautomationpartnership.co.uk
Message-ID: <40615c33@forums-2-dub>
X-Original-Trace: 24 Mar 2004 02:00:19 -0800, tap.theautomationpartnership.co.uk
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 24 Mar 2004 02:00:19 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 24 Mar 2004 02:00:22 -0800
X-Trace: forums-1-dub 1080122422 10.22.108.75 (24 Mar 2004 02:00:22 -0800)
X-Original-Trace: 24 Mar 2004 02:00:22 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2832
Article PK: 6304

OK, I did this, and you closed the bug saying it wasn't a bug! I tried to
reopen the case, and explain why this is indeeed a bug, but the status
doesn't appear to have changed from closed.

Modan

"modan" <me@here.com> wrote in message news:403b8676@forums-2-dub...
> OK, thanks
>
> "Chris Keating (iAnywhere Solutions)" <FightSpam_keating@iAnywhere.com>
> wrote in message news:403b841c$1@forums-1-dub...
> > You'll probably need to contact Technical Support to get this issue
> > resolved. Within North America, you can reach Technical Support by
> calling
> > 1-800-8SYBASE. If you are located outside of North America, you can
find
> > the contact information for your local Sybase Support Center by going to
> the
> > Sybase home page at http://www.sybase.com, clicking on the "Support"
link
> at
> > the top of the page, and then choosing "Support Centers" from the left
> hand
> > frame.
> >
> > You can always report a bug for free by visiting CaseXpress at
> > http://casexpress.sybase.com/cx/cx.stm. When submitting calls through
> > CaseXpress, please make sure to give an accurate description of the
steps
> > you take to reproduce the bug, and include any files that may be needed
> > (database, source code) to reproduce the problem.
> >
> >
> > --
> >
> > Chris Keating
> > Sybase Adaptive Server Anywhere Professional Version 8
> >
> >
>
****************************************************************************
> > *
> > Sign up today for your copy of the SQL Anywhere Studio 9 Developer
Edition
> > and try out the market-leading database for mobile, embedded and small
to
> > medium sized business environments for free!
> >
> > http://www.ianywhere.com/promos/deved/index.html
> >
> >
>
****************************************************************************
> > *
> >
> > iAnywhere Solutions http://www.iAnywhere.com
> >
> > ** Please only post to the newsgroup
> >
> > ** Whitepapers can be found at http://www.iAnywhere.com/developer
> > ** EBFs can be found at http://downloads.sybase.com/swx/sdmain.stm
> > ** Use CaseXpress to report bugs http://casexpress.sybase.com
> >
> >
>
****************************************************************************
> > *
> >
> > "modan" <me@here.com> wrote in message news:403b81b1@forums-2-dub...
> > > Things have gone a bit quiet on this thread. I guess that means no
one
> > has
> > > any ideas?
> > >
> > > FWIW this is what I think is happening.
> > >
> > > Thread 1 updates Table A
> > > Thread 1 gets interupted by Thread 2
> > > Thread 2 updates Table B
> > > Thread 2 tries to update Table A, but blocks on Thread 1's update
...and
> > > this is where things seem to go wrong.
> > > What should happen is that Thread 2 relinquishes control back to
Thread
> 1,
> > > which then tries to update Table B, and the engine detects a deadlock
> > > However, we know that this is not happening from the logging that we
> have
> > > been collecting, so presumably this means that something has happened
> > within
> > > the internals of the .Net Provider (iAnywhere) that is preventing the
> > Thread
> > > 1 from executing in later versions.
> > >
> > > Would it be useful for me to provide the source code that exhibits
this
> > > behaviour? I think it is clear that this is not a problem with our
> > > application since it worked with an earlier Sybase build, but perhaps
it
> > > would help to make it clearer where things are going wrong
> > >
> > >
> > > "Robert Waywell" <nospam_rwaywell@ianywhere.com> wrote in message
> > > news:40322308@forums-2-dub...
> > > > You can check the build number on the dboledb9.dll to see if was one
> of
> > > the
> > > > files updated in the EBF. Personally I suspect that the issue is in
> the
> > > > application rather than in the provider. One possibility would be
the
> > > > failure to handle the deadlock detected error on thread X and tell
> > thread
> > > Y
> > > > to continue.
> > > >
> > > > --
> > > > -----------------------------------------------
> > > > Robert Waywell
> > > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > > Sybase Certified Professional
> > > >
> > > > Sybase's iAnywhere Solutions
> > > >
> > > > Please respond ONLY to newsgroup
> > > >
> > > > EBF's and Patches: http://downloads.sybase.com
> > > > choose SQL Anywhere Studio >> change 'time frame' to all
> > > >
> > > > To Submit Bug Reports:
> > > >
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > > ug
> > > >
> > > > SQL Anywhere Studio Supported Platforms and Support Status
> > > > http://my.sybase.com/detail?id=1002288
> > > >
> > > > Whitepapers, TechDocs, and bug fixes are all available through the
> > > iAnywhere
> > > > Developer Community at www.ianywhere.com/developer
> > > >
> > > > <modan> wrote in message news:4030e475.1d83.846930886@sybase.com...
> > > > > Exactly!
> > > > >
> > > > > I have just added some extra tracing in the application, and
> > > > > confirmed that both threads execute the line
> > > > >
> > > > > cmd2.ExecuteNonQuery();
> > > > >
> > > > > So the problem might be related to the .Net Data Provider
> > > > > rather than the Database Engine. Does this get changed as
> > > > > part of an EBF?
> > > > >
> > > > > > So what you have observed and described is a situation in
> > > > > > which there is no deadlock in the database engine as
> > > > > > confirmed by the sa_conn_info output, however there does
> > > > > > appear to be a deadlock in the application.
> > > > > >
> > > > > > --
> > > > > > -----------------------------------------------
> > > > > > Robert Waywell
> > > > > > Sybase Adaptive Server Anywhere Developer - Version 8
> > > > > > Sybase Certified Professional
> > > > > >
> > > > > > Sybase's iAnywhere Solutions
> > > > > >
> > > > > > Please respond ONLY to newsgroup
> > > > > >
> > > > > > EBF's and Patches: http://downloads.sybase.com
> > > > > > choose SQL Anywhere Studio >> change 'time frame' to all
> > > > > >
> > > > > > To Submit Bug Reports:
> > > > > >
> > > > >
> > > >
> > >
> >
>
http://case-express.sybase.com/cx/cx.stm?starturl=casemessage.ssc?CASETYPE=B
> > > > > > ug
> > > > > >
> > > > > > SQL Anywhere Studio Supported Platforms and Support Status
> > > > > > http://my.sybase.com/detail?id=1002288
> > > > > >
> > > > > > Whitepapers, TechDocs, and bug fixes are all available
> > > > > > through the iAnywhere Developer Community at
> > > > > > www.ianywhere.com/developer
> > > > > >
> > > > > > <modan> wrote in message
> > > > > > > news:4030b328.234a.846930886@sybase.com... Just to be
> > > > > > > clear, our program invokes 2 threads. The same program
> > > > > > > can be run in a different mode where the program only
> > > > > > > runs one of the threads, allowing us to run two
> > > > > > > instances. When we run in this mode (two applications
> > > > > > rather than two threads) then the deadlock detection works
> > > > > > > , and behaves as expected.
> > > > > > >
> > > > > > > > What settings do you use for the BLOCKING and
> > > > > > > > BLOCKING_TIMEOUT options?
> > > > > > > >
> > > > > > > > Is it possible that the difference between 1250 and
> > > > > > > > 1252 is that a different connection is being chosen as
> > > > > > > > the "victim" when a deadlock is detected?
> > > > > > > >
> > > > > > > > Breck
> > > > > > > >
> > > > > > > > On 13 Feb 2004 09:12:18 -0800, modan wrote:
> > > > > > > >
> > > > > > > > >Hi, I'm going to be taking this up, as Dave is off on
> > > > > > > > >holiday.
> > > > > > > > >
> > > > > > > > >Anyway, I have confirmed that the build is 1252.
> > > > > > > > >
> > > > > > > > >In answer to your other question, sa_conn_info shows
> > > > > > > > one >blocked connection, not two.
> > > > > > > > >
> > > > > > > > >Both show ReqType of STMT_EXECUTE, and identical
> > > > > > > > >lastreqtime, and UncmtOps = 0. The "unblocked" one
> > > > > > > > shows a >LastIdle greater than the blocked one by
> > > > > > > > either 1 or 2. >
> > > > > > > > >Modan
> > > > > > > > >
> > > > > > > > >> So sa_conn_info showed two blocked connections,
> > > > > > > > blocked on >> each other? That's a situation that
> > > > > > > > should never actually >> be viewable IMHO.
> > > > > > > > >>
> > > > > > > > >> Are you sure you're on 1252, not build 1270?
> > > > > > > > >>
> > > > > > > > >> Please excuse my dumb questions, but I am chasing a
> > > > > > > > >> related issue.
> > > > > > > > >>
> > > > > > > > >> Breck
> > > > > > > > >>
> > > > > > > > >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> > > > > > > > >>
> > > > > > > > >> >I found the "Deadlock Detected" worked when the
> > > > > > > > threads >> were >in separate applications but not when
> > > > > > > > in the same >> >application
> > > > > > > > >> >
> > > > > > > > >> >It did show a lock Cycle, Only 3 out of the 4
> > > > > > > > Updates to >> the >database were processed by the
> > > > > > > > server. >> >
> > > > > > > > >> >Dave
> > > > > > > > >> >
> > > > > > > > >> >
> > > > > > > > >> >> FWIW I just ran a deadlock test using two
> > > > > > > > connections >> and >> got "Deadlock detected" as
> > > > > > > > expected, in build 1252. >> >>
> > > > > > > > >> >> What does sa_conn_info show when you run it on a
> > > > > > > > 3rd >> >> connection? Does it show an actual lock
> > > > > > > > cycle >> (deadlock)? >>
> > > > > > > > >> >> Breck
> > > > > > > > >> >>
> > > > > > > > >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage
> > > > > > > > wrote: >> >>
> > > > > > > > >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> > > > > > > > Build >> 1108 >> >
> > > > > > > > >> >> >When Two Threads created a DeadLock, I get the
> > > > > > > > Correct >> >> >Deadlock Error returned for one of the
> > > > > > > > Threads, >> >> >
> > > > > > > > >> >> >However when I install the Express Bug Fix,
> > > > > > > > Update of >> >> 9.0.0 >to Build 1252
> > > > > > > > >> >> >
> > > > > > > > >> >> >I no longer get the DeadLock Error, The ASA
> > > > > > > > Commands >> just >> >Hang.
> > > > > > > > >> >> >
> > > > > > > > >> >> >Has anyone else seen this ?
> > > > > > > > >> >> >
> > > > > > > > >> >> >Thanks in advance
> > > > > > > > >> >> >
> > > > > > > > >> >> >Dave
> > > > > > > > >> >>
> > > > > > > > >> >> --
> > > > > > > > >> >> bcarter@risingroad.com
> > > > > > > > >> >> Mobile and Distributed Enterprise Database
> > > > > > > > Applications >> >> www.risingroad.com
> > > > > > > > >>
> > > > > > > > >> --
> > > > > > > > >> bcarter@risingroad.com
> > > > > > > > >> Mobile and Distributed Enterprise Database
> > > > > > > > Applications >> www.risingroad.com
> > > > > > > >
> > > > > > > > --
> > > > > > > > bcarter@risingroad.com
> > > > > > > > Mobile and Distributed Enterprise Database
> > > > > > > > Applications www.risingroad.com
> > > > > >
> > > > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>


Stephen Rice Posted on 2004-03-25 02:57:26.0Z
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
From: Stephen Rice <srice_nospam@ianywhere.com>
References: <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub> <403b81b1@forums-2-dub> <403b841c$1@forums-1-dub> <403b8676@forums-2-dub> <40615c33@forums-2-dub>
Organization: iAnywhere Solutions
Message-ID: <Xns94B6DEDA88012sricenospamianywhere@192.138.151.106>
User-Agent: Xnews/5.04.25
X-Original-NNTP-Posting-Host: toronto-hse-ppp3932066.sympatico.ca
X-Original-Trace: 24 Mar 2004 18:57:22 -0800, toronto-hse-ppp3932066.sympatico.ca
Lines: 25
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 24 Mar 2004 18:57:24 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 24 Mar 2004 18:57:26 -0800
X-Trace: forums-1-dub 1080183446 10.22.108.75 (24 Mar 2004 18:57:26 -0800)
X-Original-Trace: 24 Mar 2004 18:57:26 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2844
Article PK: 6314


"modan" <me@here.com> wrote in news:40615c33@forums-2-dub:

> OK, I did this, and you closed the bug saying it wasn't a bug! I
> tried to reopen the case, and explain why this is indeeed a bug, but
> the status doesn't appear to have changed from closed.
>
> Modan
>
>

What was your case number?

/steve


--
Stephen Rice
Technical Services Manager
iAnywhere Solutions

- Please Post-
Whitepapers, TechDocs, and bug fixes are all available through the
iAnywhere
Developer Community at www.ianywhere.com/developer


modan Posted on 2004-03-25 12:30:54.0Z
From: "modan" <me@here.com>
Newsgroups: ianywhere.public.general
References: <4030e475.1d83.846930886@sybase.com> <40322308@forums-2-dub> <403b81b1@forums-2-dub> <403b841c$1@forums-1-dub> <403b8676@forums-2-dub> <40615c33@forums-2-dub> <Xns94B6DEDA88012sricenospamianywhere@192.138.151.106>
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Lines: 32
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1158
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
X-Original-NNTP-Posting-Host: tap.theautomationpartnership.co.uk
Message-ID: <4062d0fa@forums-2-dub>
X-Original-Trace: 25 Mar 2004 04:30:50 -0800, tap.theautomationpartnership.co.uk
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 25 Mar 2004 04:30:51 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 25 Mar 2004 04:30:54 -0800
X-Trace: forums-1-dub 1080217854 10.22.108.75 (25 Mar 2004 04:30:54 -0800)
X-Original-Trace: 25 Mar 2004 04:30:54 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2845
Article PK: 6315

The case number is 20340071

"Stephen Rice" <srice_nospam@ianywhere.com> wrote in message
news:Xns94B6DEDA88012sricenospamianywhere@192.138.151.106...
> "modan" <me@here.com> wrote in news:40615c33@forums-2-dub:
>
> > OK, I did this, and you closed the bug saying it wasn't a bug! I
> > tried to reopen the case, and explain why this is indeeed a bug, but
> > the status doesn't appear to have changed from closed.
> >
> > Modan
> >
> >
>
> What was your case number?
>
> /steve
>
>
> --
> Stephen Rice
> Technical Services Manager
> iAnywhere Solutions
>
> - Please Post-
> Whitepapers, TechDocs, and bug fixes are all available through the
> iAnywhere
> Developer Community at www.ianywhere.com/developer
>


modan Posted on 2004-02-16 11:29:43.0Z
Sender: 2204.4030a906.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <4030a9a7.2209.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><fp9q20tdgtkmh41l6tsb5d7fgm9c35945s@4ax.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 16 Feb 2004 03:29:43 -0800
X-Trace: forums-1-dub 1076930983 10.22.241.41 (16 Feb 2004 03:29:43 -0800)
X-Original-Trace: 16 Feb 2004 03:29:43 -0800, 10.22.241.41
Lines: 95
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2509
Article PK: 5649

BLOCKING = ON
BLOCKING_TIMEOUT = 0

Program is simple test App, so should not care which thread
is rolled back/terminated/whatever.

> What settings do you use for the BLOCKING and
> BLOCKING_TIMEOUT options?
>
> Is it possible that the difference between 1250 and 1252
> is that a different connection is being chosen as the
> "victim" when a deadlock is detected?
>
> Breck
>
> On 13 Feb 2004 09:12:18 -0800, modan wrote:
>
> >Hi, I'm going to be taking this up, as Dave is off on
> >holiday.
> >
> >Anyway, I have confirmed that the build is 1252.
> >
> >In answer to your other question, sa_conn_info shows one
> >blocked connection, not two.
> >
> >Both show ReqType of STMT_EXECUTE, and identical
> >lastreqtime, and UncmtOps = 0. The "unblocked" one shows
> a >LastIdle greater than the blocked one by either 1 or 2.
> >
> >Modan
> >
> >> So sa_conn_info showed two blocked connections, blocked
> on >> each other? That's a situation that should never
> actually >> be viewable IMHO.
> >>
> >> Are you sure you're on 1252, not build 1270?
> >>
> >> Please excuse my dumb questions, but I am chasing a
> >> related issue.
> >>
> >> Breck
> >>
> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> >>
> >> >I found the "Deadlock Detected" worked when the
> threads >> were >in separate applications but not when in
> the same >> >application
> >> >
> >> >It did show a lock Cycle, Only 3 out of the 4 Updates
> to >> the >database were processed by the server.
> >> >
> >> >Dave
> >> >
> >> >
> >> >> FWIW I just ran a deadlock test using two
> connections >> and >> got "Deadlock detected" as expected,
> in build 1252. >> >>
> >> >> What does sa_conn_info show when you run it on a 3rd
> >> >> connection? Does it show an actual lock cycle
> >> (deadlock)? >>
> >> >> Breck
> >> >>
> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> >> >>
> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> Build >> 1108 >> >
> >> >> >When Two Threads created a DeadLock, I get the
> Correct >> >> >Deadlock Error returned for one of the
> Threads, >> >> >
> >> >> >However when I install the Express Bug Fix, Update
> of >> >> 9.0.0 >to Build 1252
> >> >> >
> >> >> >I no longer get the DeadLock Error, The ASA
> Commands >> just >> >Hang.
> >> >> >
> >> >> >Has anyone else seen this ?
> >> >> >
> >> >> >Thanks in advance
> >> >> >
> >> >> >Dave
> >> >>
> >> >> --
> >> >> bcarter@risingroad.com
> >> >> Mobile and Distributed Enterprise Database
> Applications >> >> www.risingroad.com
> >>
> >> --
> >> bcarter@risingroad.com
> >> Mobile and Distributed Enterprise Database Applications
> >> www.risingroad.com
>
> --
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-13 19:39:22.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <0i8q209vqent7vl7jgo268kjiiv6u57m4e@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
X-Original-Trace: 13 Feb 2004 11:50:25 -0800, bcarter.sentex.ca
Lines: 96
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 11:21:11 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 11:39:22 -0800
X-Trace: forums-1-dub 1076701162 10.22.108.75 (13 Feb 2004 11:39:22 -0800)
X-Original-Trace: 13 Feb 2004 11:39:22 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2495
Article PK: 5639

If there is only one blocked connection then it isn't a deadlock.

Which means either (a) there never was a deadlock or (b) there was a
deadlock and it got resolved. Can you turn on request level logging to
see what the sequence of commands hitting the server are?

-- On...
CALL sa_server_option ( 'Request_level_log_file', 'r.txt' );
CALL sa_server_option ( 'Request_level_logging', 'SQL+hostvars' );

-- Off...
--CALL sa_server_option ( 'Request_level_logging', 'NONE' );

Breck

On 13 Feb 2004 09:12:18 -0800, modan wrote:

>Hi, I'm going to be taking this up, as Dave is off on
>holiday.
>
>Anyway, I have confirmed that the build is 1252.
>
>In answer to your other question, sa_conn_info shows one
>blocked connection, not two.
>
>Both show ReqType of STMT_EXECUTE, and identical
>lastreqtime, and UncmtOps = 0. The "unblocked" one shows a
>LastIdle greater than the blocked one by either 1 or 2.
>
>Modan
>
>> So sa_conn_info showed two blocked connections, blocked on
>> each other? That's a situation that should never actually
>> be viewable IMHO.
>>
>> Are you sure you're on 1252, not build 1270?
>>
>> Please excuse my dumb questions, but I am chasing a
>> related issue.
>>
>> Breck
>>
>> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>>
>> >I found the "Deadlock Detected" worked when the threads
>> were >in separate applications but not when in the same
>> >application
>> >
>> >It did show a lock Cycle, Only 3 out of the 4 Updates to
>> the >database were processed by the server.
>> >
>> >Dave
>> >
>> >
>> >> FWIW I just ran a deadlock test using two connections
>> and >> got "Deadlock detected" as expected, in build 1252.
>> >>
>> >> What does sa_conn_info show when you run it on a 3rd
>> >> connection? Does it show an actual lock cycle
>> (deadlock)? >>
>> >> Breck
>> >>
>> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>> >>
>> >> >Working with Sybase SQL Anywhere Verion 9.0.0. Build
>> 1108 >> >
>> >> >When Two Threads created a DeadLock, I get the Correct
>> >> >Deadlock Error returned for one of the Threads,
>> >> >
>> >> >However when I install the Express Bug Fix, Update of
>> >> 9.0.0 >to Build 1252
>> >> >
>> >> >I no longer get the DeadLock Error, The ASA Commands
>> just >> >Hang.
>> >> >
>> >> >Has anyone else seen this ?
>> >> >
>> >> >Thanks in advance
>> >> >
>> >> >Dave
>> >>
>> >> --
>> >> bcarter@risingroad.com
>> >> Mobile and Distributed Enterprise Database Applications
>> >> www.risingroad.com
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


modan Posted on 2004-02-16 11:26:53.0Z
Sender: 2201.4030a8ea.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <4030a8fd.2203.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><0i8q209vqent7vl7jgo268kjiiv6u57m4e@4ax.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 16 Feb 2004 03:26:53 -0800
X-Trace: forums-1-dub 1076930813 10.22.241.41 (16 Feb 2004 03:26:53 -0800)
X-Original-Trace: 16 Feb 2004 03:26:53 -0800, 10.22.241.41
Lines: 215
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2508
Article PK: 5646

Actually I am pretty sure there is a deadlock.

The program creates two threads.

First thread does

update A followed by update B

Second thread does

update B followed by update A

3 of the updates happen, according to the log, and both
threads end up blocked (only one is blocked according to
sa_conn_info though.

It may be worth mentioning that this test app is written in
C# and uses the .Net Data Provider.

Request_level_logging is as follows:-

02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
02/16 10:54:11.085 ** CONNECT conn: 304771280 UID=DBA
02/16 10:54:11.085 ** DONE conn: 304771280 CONNECT
Conn=1744647885
02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_PREPARE
"select db_property('charset')"
02/16 10:54:11.086 ** DONE conn: 304771280 STMT_PREPARE
Stmt=65536
02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.086 ** DONE conn: 304771280 STMT_EXECUTE
02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_DROP
Stmt=65536
02/16 10:54:11.086 ** DONE conn: 304771280 STMT_DROP
02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_PREPARE
"select connection_property( 'max_statement_count'
)"
02/16 10:54:11.086 ** DONE conn: 304771280 STMT_PREPARE
Stmt=65537
02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.086 ** DONE conn: 304771280 STMT_EXECUTE
02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_DROP
Stmt=65537
02/16 10:54:11.088 ** DONE conn: 304771280 STMT_DROP
02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_PREPARE
"SET TEMPORARY OPTION ISOLATION_LEVEL=1"
02/16 10:54:11.088 ** DONE conn: 304771280 STMT_PREPARE
Stmt=65538
02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.089 ** DONE conn: 304771280 STMT_EXECUTE
02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_DROP
Stmt=65538
02/16 10:54:11.096 ** DONE conn: 304771280 STMT_DROP
02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_PREPARE
"UPDATE TableA SET A = 2"
02/16 10:54:11.096 ** DONE conn: 304771280 STMT_PREPARE
Stmt=65539
02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.096 ** DONE conn: 304771280 STMT_EXECUTE
02/16 10:54:11.098 ** REQUEST conn: 304761224 CONNECT
02/16 10:54:11.098 ** CONNECT conn: 304761224 UID=DBA
02/16 10:54:11.098 ** DONE conn: 304761224 CONNECT
Conn=167015670
02/16 10:54:11.098 ** REQUEST conn: 304761224 STMT_PREPARE
"select db_property('charset')"
02/16 10:54:11.098 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65536
02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.099 ** DONE conn: 304761224 STMT_EXECUTE
02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_DROP
Stmt=65536
02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_PREPARE
"select connection_property( 'max_statement_count'
)"
02/16 10:54:11.099 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65537
02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.099 ** DONE conn: 304761224 STMT_EXECUTE
02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_DROP
Stmt=65537
02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
"SET TEMPORARY OPTION ISOLATION_LEVEL=1"
02/16 10:54:11.100 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65538
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.100 ** DONE conn: 304761224 STMT_EXECUTE
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_DROP
Stmt=65538
02/16 10:54:11.100 ** DONE conn: 304761224 STMT_DROP
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
"UPDATE TableB SET B = 2"
02/16 10:54:11.101 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65539
02/16 10:54:11.101 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1
02/16 10:54:11.101 ** DONE conn: 304761224 STMT_EXECUTE
02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_DROP
Stmt=65539
02/16 10:54:11.102 ** DONE conn: 304761224 STMT_DROP
02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_PREPARE
"UPDATE TableA SET A = 2"
02/16 10:54:11.102 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65540
02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1

> If there is only one blocked connection then it isn't a
> deadlock.
>
> Which means either (a) there never was a deadlock or (b)
> there was a deadlock and it got resolved. Can you turn on
> request level logging to see what the sequence of commands
> hitting the server are?
>
> -- On...
> CALL sa_server_option ( 'Request_level_log_file', 'r.txt'
> ); CALL sa_server_option ( 'Request_level_logging',
> 'SQL+hostvars' );
>
> -- Off...
> --CALL sa_server_option ( 'Request_level_logging', 'NONE'
> );
>
> Breck
>
>
> On 13 Feb 2004 09:12:18 -0800, modan wrote:
>
> >Hi, I'm going to be taking this up, as Dave is off on
> >holiday.
> >
> >Anyway, I have confirmed that the build is 1252.
> >
> >In answer to your other question, sa_conn_info shows one
> >blocked connection, not two.
> >
> >Both show ReqType of STMT_EXECUTE, and identical
> >lastreqtime, and UncmtOps = 0. The "unblocked" one shows
> a >LastIdle greater than the blocked one by either 1 or 2.
> >
> >Modan
> >
> >> So sa_conn_info showed two blocked connections, blocked
> on >> each other? That's a situation that should never
> actually >> be viewable IMHO.
> >>
> >> Are you sure you're on 1252, not build 1270?
> >>
> >> Please excuse my dumb questions, but I am chasing a
> >> related issue.
> >>
> >> Breck
> >>
> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> >>
> >> >I found the "Deadlock Detected" worked when the
> threads >> were >in separate applications but not when in
> the same >> >application
> >> >
> >> >It did show a lock Cycle, Only 3 out of the 4 Updates
> to >> the >database were processed by the server.
> >> >
> >> >Dave
> >> >
> >> >
> >> >> FWIW I just ran a deadlock test using two
> connections >> and >> got "Deadlock detected" as expected,
> in build 1252. >> >>
> >> >> What does sa_conn_info show when you run it on a 3rd
> >> >> connection? Does it show an actual lock cycle
> >> (deadlock)? >>
> >> >> Breck
> >> >>
> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> >> >>
> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> Build >> 1108 >> >
> >> >> >When Two Threads created a DeadLock, I get the
> Correct >> >> >Deadlock Error returned for one of the
> Threads, >> >> >
> >> >> >However when I install the Express Bug Fix, Update
> of >> >> 9.0.0 >to Build 1252
> >> >> >
> >> >> >I no longer get the DeadLock Error, The ASA
> Commands >> just >> >Hang.
> >> >> >
> >> >> >Has anyone else seen this ?
> >> >> >
> >> >> >Thanks in advance
> >> >> >
> >> >> >Dave
> >> >>
> >> >> --
> >> >> bcarter@risingroad.com
> >> >> Mobile and Distributed Enterprise Database
> Applications >> >> www.risingroad.com
> >>
> >> --
> >> bcarter@risingroad.com
> >> Mobile and Distributed Enterprise Database Applications
> >> www.risingroad.com
>
> --
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-16 12:56:05.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <nfd1305ajpto3ab4tlj0pg66aqk7mkhp3c@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><0i8q209vqent7vl7jgo268kjiiv6u57m4e@4ax.com> <4030a8fd.2203.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: bcarter.sentex.ca
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
Date: 16 Feb 2004 04:56:05 -0800
X-Trace: forums-1-dub 1076936165 64.7.134.118 (16 Feb 2004 04:56:05 -0800)
X-Original-Trace: 16 Feb 2004 04:56:05 -0800, bcarter.sentex.ca
Lines: 277
X-Authenticated-User: TeamPS
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2511
Article PK: 5648

No, as far as I can tell only 2 of the updates finish, and the third
one is blocked (STMT_EXECUTE REQUEST but no DONE).

Also, I do not see any attempt by the first connection to do its
second update. Here is the trimmed-down request level log, with
comments...

===== Connection 1 completes UPDATE TableA...
02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_PREPARE
"SET TEMPORARY OPTION ISOLATION_LEVEL=1"
02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_PREPARE
"UPDATE TableA SET A = 2"
02/16 10:54:11.096 ** DONE conn: 304771280 STMT_EXECUTE

===== Connection 1 completes UPDATE TableB, blocked on UPDATE
TableA...
02/16 10:54:11.098 ** REQUEST conn: 304761224 CONNECT
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
"SET TEMPORARY OPTION ISOLATION_LEVEL=1"
02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
"UPDATE TableB SET B = 2"
02/16 10:54:11.101 ** DONE conn: 304761224 STMT_EXECUTE

02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_PREPARE
"UPDATE TableA SET A = 2"
02/16 10:54:11.102 ** DONE conn: 304761224 STMT_PREPARE
Stmt=65540
02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_EXECUTE
Stmt=-1

Is there nothing happening past this point? If not, there is no
deadlock because the first connection (A.K.A. 304771280) does not
attempt to update TableB. What is shown here is a simple block: the
second connection (A.K.A. 304761224) is waiting for the first
connection to release the lock(s) on TableA.

Put it another way, you need to see 4 updates for there to be a
deadlock, and only 3 are shown.

FWIW you can see more details of what is happening two ways. First,
CALL sa_locks() to see which connection has locks on which table.
Second, you can see the transaction log in readable format by running
this command (no need to shut down the server first):

"%ASANY9%\win32\dbtran.exe" -a -d -c "whatever" -s -y whatever.sql

The -f option is handy to reduce the amount of output, especially if
you do an explicit CHECKPOINT just before starting your test run.

Breck

On 16 Feb 2004 03:26:53 -0800, modan wrote:

>Actually I am pretty sure there is a deadlock.
>
>The program creates two threads.
>
>First thread does
>
>update A followed by update B
>
>Second thread does
>
>update B followed by update A
>
>3 of the updates happen, according to the log, and both
>threads end up blocked (only one is blocked according to
>sa_conn_info though.
>
>It may be worth mentioning that this test app is written in
>C# and uses the .Net Data Provider.
>
>Request_level_logging is as follows:-
>
>02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
>02/16 10:54:11.085 ** CONNECT conn: 304771280 UID=DBA
>02/16 10:54:11.085 ** DONE conn: 304771280 CONNECT
> Conn=1744647885
>02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_PREPARE
> "select db_property('charset')"
>02/16 10:54:11.086 ** DONE conn: 304771280 STMT_PREPARE
> Stmt=65536
>02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.086 ** DONE conn: 304771280 STMT_EXECUTE
>02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_DROP
> Stmt=65536
>02/16 10:54:11.086 ** DONE conn: 304771280 STMT_DROP
>02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_PREPARE
> "select connection_property( 'max_statement_count'
>)"
>02/16 10:54:11.086 ** DONE conn: 304771280 STMT_PREPARE
> Stmt=65537
>02/16 10:54:11.086 ** REQUEST conn: 304771280 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.086 ** DONE conn: 304771280 STMT_EXECUTE
>02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_DROP
> Stmt=65537
>02/16 10:54:11.088 ** DONE conn: 304771280 STMT_DROP
>02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_PREPARE
> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
>02/16 10:54:11.088 ** DONE conn: 304771280 STMT_PREPARE
> Stmt=65538
>02/16 10:54:11.088 ** REQUEST conn: 304771280 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.089 ** DONE conn: 304771280 STMT_EXECUTE
>02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_DROP
> Stmt=65538
>02/16 10:54:11.096 ** DONE conn: 304771280 STMT_DROP
>02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_PREPARE
> "UPDATE TableA SET A = 2"
>02/16 10:54:11.096 ** DONE conn: 304771280 STMT_PREPARE
> Stmt=65539
>02/16 10:54:11.096 ** REQUEST conn: 304771280 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.096 ** DONE conn: 304771280 STMT_EXECUTE
>02/16 10:54:11.098 ** REQUEST conn: 304761224 CONNECT
>02/16 10:54:11.098 ** CONNECT conn: 304761224 UID=DBA
>02/16 10:54:11.098 ** DONE conn: 304761224 CONNECT
> Conn=167015670
>02/16 10:54:11.098 ** REQUEST conn: 304761224 STMT_PREPARE
> "select db_property('charset')"
>02/16 10:54:11.098 ** DONE conn: 304761224 STMT_PREPARE
> Stmt=65536
>02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.099 ** DONE conn: 304761224 STMT_EXECUTE
>02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_DROP
> Stmt=65536
>02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
>02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_PREPARE
> "select connection_property( 'max_statement_count'
>)"
>02/16 10:54:11.099 ** DONE conn: 304761224 STMT_PREPARE
> Stmt=65537
>02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.099 ** DONE conn: 304761224 STMT_EXECUTE
>02/16 10:54:11.099 ** REQUEST conn: 304761224 STMT_DROP
> Stmt=65537
>02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
>02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
>02/16 10:54:11.100 ** DONE conn: 304761224 STMT_PREPARE
> Stmt=65538
>02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.100 ** DONE conn: 304761224 STMT_EXECUTE
>02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_DROP
> Stmt=65538
>02/16 10:54:11.100 ** DONE conn: 304761224 STMT_DROP
>02/16 10:54:11.100 ** REQUEST conn: 304761224 STMT_PREPARE
> "UPDATE TableB SET B = 2"
>02/16 10:54:11.101 ** DONE conn: 304761224 STMT_PREPARE
> Stmt=65539
>02/16 10:54:11.101 ** REQUEST conn: 304761224 STMT_EXECUTE
> Stmt=-1
>02/16 10:54:11.101 ** DONE conn: 304761224 STMT_EXECUTE
>02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_DROP
> Stmt=65539
>02/16 10:54:11.102 ** DONE conn: 304761224 STMT_DROP
>02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_PREPARE
> "UPDATE TableA SET A = 2"
>02/16 10:54:11.102 ** DONE conn: 304761224 STMT_PREPARE
> Stmt=65540
>02/16 10:54:11.102 ** REQUEST conn: 304761224 STMT_EXECUTE
> Stmt=-1
>
>> If there is only one blocked connection then it isn't a
>> deadlock.
>>
>> Which means either (a) there never was a deadlock or (b)
>> there was a deadlock and it got resolved. Can you turn on
>> request level logging to see what the sequence of commands
>> hitting the server are?
>>
>> -- On...
>> CALL sa_server_option ( 'Request_level_log_file', 'r.txt'
>> ); CALL sa_server_option ( 'Request_level_logging',
>> 'SQL+hostvars' );
>>
>> -- Off...
>> --CALL sa_server_option ( 'Request_level_logging', 'NONE'
>> );
>>
>> Breck
>>
>>
>> On 13 Feb 2004 09:12:18 -0800, modan wrote:
>>
>> >Hi, I'm going to be taking this up, as Dave is off on
>> >holiday.
>> >
>> >Anyway, I have confirmed that the build is 1252.
>> >
>> >In answer to your other question, sa_conn_info shows one
>> >blocked connection, not two.
>> >
>> >Both show ReqType of STMT_EXECUTE, and identical
>> >lastreqtime, and UncmtOps = 0. The "unblocked" one shows
>> a >LastIdle greater than the blocked one by either 1 or 2.
>> >
>> >Modan
>> >
>> >> So sa_conn_info showed two blocked connections, blocked
>> on >> each other? That's a situation that should never
>> actually >> be viewable IMHO.
>> >>
>> >> Are you sure you're on 1252, not build 1270?
>> >>
>> >> Please excuse my dumb questions, but I am chasing a
>> >> related issue.
>> >>
>> >> Breck
>> >>
>> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>> >>
>> >> >I found the "Deadlock Detected" worked when the
>> threads >> were >in separate applications but not when in
>> the same >> >application
>> >> >
>> >> >It did show a lock Cycle, Only 3 out of the 4 Updates
>> to >> the >database were processed by the server.
>> >> >
>> >> >Dave
>> >> >
>> >> >
>> >> >> FWIW I just ran a deadlock test using two
>> connections >> and >> got "Deadlock detected" as expected,
>> in build 1252. >> >>
>> >> >> What does sa_conn_info show when you run it on a 3rd
>> >> >> connection? Does it show an actual lock cycle
>> >> (deadlock)? >>
>> >> >> Breck
>> >> >>
>> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>> >> >>
>> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
>> Build >> 1108 >> >
>> >> >> >When Two Threads created a DeadLock, I get the
>> Correct >> >> >Deadlock Error returned for one of the
>> Threads, >> >> >
>> >> >> >However when I install the Express Bug Fix, Update
>> of >> >> 9.0.0 >to Build 1252
>> >> >> >
>> >> >> >I no longer get the DeadLock Error, The ASA
>> Commands >> just >> >Hang.
>> >> >> >
>> >> >> >Has anyone else seen this ?
>> >> >> >
>> >> >> >Thanks in advance
>> >> >> >
>> >> >> >Dave
>> >> >>
>> >> >> --
>> >> >> bcarter@risingroad.com
>> >> >> Mobile and Distributed Enterprise Database
>> Applications >> >> www.risingroad.com
>> >>
>> >> --
>> >> bcarter@risingroad.com
>> >> Mobile and Distributed Enterprise Database Applications
>> >> www.risingroad.com
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
SQL Anywhere Studio 9 Developer's Guide
ISBN 1-55622-506-7
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


modan Posted on 2004-02-16 16:58:05.0Z
Sender: 1eab.4030f3f1.1804289383@sybase.com
From: modan
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
X-Mailer: WebNews to Mail Gateway v1.1s
Message-ID: <4030f93f.1f22.846930886@sybase.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><0i8q209vqent7vl7jgo268kjiiv6u57m4e@4ax.com> <4030a8fd.2203.846930886@sybase.com><nfd1305ajpto3ab4tlj0pg66aqk7mkhp3c@4ax.com>
X-Original-NNTP-Posting-Host: 10.22.241.42
X-Original-Trace: 16 Feb 2004 09:09:19 -0800, 10.22.241.42
Lines: 297
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 16 Feb 2004 08:39:37 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 16 Feb 2004 08:58:05 -0800
X-Trace: forums-1-dub 1076950685 10.22.108.75 (16 Feb 2004 08:58:05 -0800)
X-Original-Trace: 16 Feb 2004 08:58:05 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2515
Article PK: 5653

As expected, translating the transaction log again shows no
sign of the second thread's query actually making it as far
as the DB engine, so it is looking increasingly likely that
the bug is in the Adaptive Server Anywhere .NET Data
Provider.

> No, as far as I can tell only 2 of the updates finish, and
> the third one is blocked (STMT_EXECUTE REQUEST but no
> DONE).
>
> Also, I do not see any attempt by the first connection to
> do its second update. Here is the trimmed-down request
> level log, with comments...
>
> ===== Connection 1 completes UPDATE TableA...
> 02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
> 02/16 10:54:11.088 ** REQUEST conn: 304771280
> STMT_PREPARE
> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
> 02/16 10:54:11.096 ** REQUEST conn: 304771280
> STMT_PREPARE
> "UPDATE TableA SET A = 2"
> 02/16 10:54:11.096 ** DONE conn: 304771280
> STMT_EXECUTE
>
> ===== Connection 1 completes UPDATE TableB, blocked on
> UPDATE TableA...
> 02/16 10:54:11.098 ** REQUEST conn: 304761224 CONNECT
> 02/16 10:54:11.100 ** REQUEST conn: 304761224
> STMT_PREPARE
> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
> 02/16 10:54:11.100 ** REQUEST conn: 304761224
> STMT_PREPARE
> "UPDATE TableB SET B = 2"
> 02/16 10:54:11.101 ** DONE conn: 304761224
> STMT_EXECUTE
>
> 02/16 10:54:11.102 ** REQUEST conn: 304761224
> STMT_PREPARE
> "UPDATE TableA SET A = 2"
> 02/16 10:54:11.102 ** DONE conn: 304761224
> STMT_PREPARE
> Stmt=65540
> 02/16 10:54:11.102 ** REQUEST conn: 304761224
> STMT_EXECUTE
> Stmt=-1
>
> Is there nothing happening past this point? If not, there
> is no deadlock because the first connection (A.K.A.
> 304771280) does not attempt to update TableB. What is
> shown here is a simple block: the second connection
> (A.K.A. 304761224) is waiting for the first connection to
> release the lock(s) on TableA.
>
> Put it another way, you need to see 4 updates for there to
> be a deadlock, and only 3 are shown.
>
> FWIW you can see more details of what is happening two
> ways. First, CALL sa_locks() to see which connection has
> locks on which table. Second, you can see the transaction
> log in readable format by running this command (no need to
> shut down the server first):
>
> "%ASANY9%\win32\dbtran.exe" -a -d -c "whatever" -s -y
> whatever.sql
>
> The -f option is handy to reduce the amount of output,
> especially if you do an explicit CHECKPOINT just before
> starting your test run.
>
> Breck
>
>
> On 16 Feb 2004 03:26:53 -0800, modan wrote:
>
> >Actually I am pretty sure there is a deadlock.
> >
> >The program creates two threads.
> >
> >First thread does
> >
> >update A followed by update B
> >
> >Second thread does
> >
> >update B followed by update A
> >
> >3 of the updates happen, according to the log, and both
> >threads end up blocked (only one is blocked according to
> >sa_conn_info though.
> >
> >It may be worth mentioning that this test app is written
> in >C# and uses the .Net Data Provider.
> >
> >Request_level_logging is as follows:-
> >
> >02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
> >02/16 10:54:11.085 ** CONNECT conn: 304771280 UID=DBA
> >02/16 10:54:11.085 ** DONE conn: 304771280 CONNECT
> > Conn=1744647885
> >02/16 10:54:11.086 ** REQUEST conn: 304771280
> > STMT_PREPARE "select db_property('charset')"
> >02/16 10:54:11.086 ** DONE conn: 304771280
> > STMT_PREPARE Stmt=65536
> >02/16 10:54:11.086 ** REQUEST conn: 304771280
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.086 ** DONE conn: 304771280
> STMT_EXECUTE >02/16 10:54:11.086 ** REQUEST conn:
> > 304771280 STMT_DROP Stmt=65536
> >02/16 10:54:11.086 ** DONE conn: 304771280 STMT_DROP
> >02/16 10:54:11.086 ** REQUEST conn: 304771280
> > STMT_PREPARE "select connection_property(
> 'max_statement_count' >)"
> >02/16 10:54:11.086 ** DONE conn: 304771280
> > STMT_PREPARE Stmt=65537
> >02/16 10:54:11.086 ** REQUEST conn: 304771280
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.086 ** DONE conn: 304771280
> STMT_EXECUTE >02/16 10:54:11.088 ** REQUEST conn:
> > 304771280 STMT_DROP Stmt=65537
> >02/16 10:54:11.088 ** DONE conn: 304771280 STMT_DROP
> >02/16 10:54:11.088 ** REQUEST conn: 304771280
> > STMT_PREPARE "SET TEMPORARY OPTION
> ISOLATION_LEVEL=1" >02/16 10:54:11.088 ** DONE conn:
> > 304771280 STMT_PREPARE Stmt=65538
> >02/16 10:54:11.088 ** REQUEST conn: 304771280
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.089 ** DONE conn: 304771280
> STMT_EXECUTE >02/16 10:54:11.096 ** REQUEST conn:
> > 304771280 STMT_DROP Stmt=65538
> >02/16 10:54:11.096 ** DONE conn: 304771280 STMT_DROP
> >02/16 10:54:11.096 ** REQUEST conn: 304771280
> > STMT_PREPARE "UPDATE TableA SET A = 2"
> >02/16 10:54:11.096 ** DONE conn: 304771280
> > STMT_PREPARE Stmt=65539
> >02/16 10:54:11.096 ** REQUEST conn: 304771280
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.096 ** DONE conn: 304771280
> STMT_EXECUTE >02/16 10:54:11.098 ** REQUEST conn:
> 304761224 CONNECT >02/16 10:54:11.098 ** CONNECT conn:
> 304761224 UID=DBA >02/16 10:54:11.098 ** DONE conn:
> > 304761224 CONNECT Conn=167015670
> >02/16 10:54:11.098 ** REQUEST conn: 304761224
> > STMT_PREPARE "select db_property('charset')"
> >02/16 10:54:11.098 ** DONE conn: 304761224
> > STMT_PREPARE Stmt=65536
> >02/16 10:54:11.099 ** REQUEST conn: 304761224
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.099 ** DONE conn: 304761224
> STMT_EXECUTE >02/16 10:54:11.099 ** REQUEST conn:
> > 304761224 STMT_DROP Stmt=65536
> >02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
> >02/16 10:54:11.099 ** REQUEST conn: 304761224
> > STMT_PREPARE "select connection_property(
> 'max_statement_count' >)"
> >02/16 10:54:11.099 ** DONE conn: 304761224
> > STMT_PREPARE Stmt=65537
> >02/16 10:54:11.099 ** REQUEST conn: 304761224
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.099 ** DONE conn: 304761224
> STMT_EXECUTE >02/16 10:54:11.099 ** REQUEST conn:
> > 304761224 STMT_DROP Stmt=65537
> >02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
> >02/16 10:54:11.100 ** REQUEST conn: 304761224
> > STMT_PREPARE "SET TEMPORARY OPTION
> ISOLATION_LEVEL=1" >02/16 10:54:11.100 ** DONE conn:
> > 304761224 STMT_PREPARE Stmt=65538
> >02/16 10:54:11.100 ** REQUEST conn: 304761224
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.100 ** DONE conn: 304761224
> STMT_EXECUTE >02/16 10:54:11.100 ** REQUEST conn:
> > 304761224 STMT_DROP Stmt=65538
> >02/16 10:54:11.100 ** DONE conn: 304761224 STMT_DROP
> >02/16 10:54:11.100 ** REQUEST conn: 304761224
> > STMT_PREPARE "UPDATE TableB SET B = 2"
> >02/16 10:54:11.101 ** DONE conn: 304761224
> > STMT_PREPARE Stmt=65539
> >02/16 10:54:11.101 ** REQUEST conn: 304761224
> > STMT_EXECUTE Stmt=-1
> >02/16 10:54:11.101 ** DONE conn: 304761224
> STMT_EXECUTE >02/16 10:54:11.102 ** REQUEST conn:
> > 304761224 STMT_DROP Stmt=65539
> >02/16 10:54:11.102 ** DONE conn: 304761224 STMT_DROP
> >02/16 10:54:11.102 ** REQUEST conn: 304761224
> > STMT_PREPARE "UPDATE TableA SET A = 2"
> >02/16 10:54:11.102 ** DONE conn: 304761224
> > STMT_PREPARE Stmt=65540
> >02/16 10:54:11.102 ** REQUEST conn: 304761224
> > STMT_EXECUTE Stmt=-1
> >
> >> If there is only one blocked connection then it isn't a
> >> deadlock.
> >>
> >> Which means either (a) there never was a deadlock or
> (b) >> there was a deadlock and it got resolved. Can you
> turn on >> request level logging to see what the sequence
> of commands >> hitting the server are?
> >>
> >> -- On...
> >> CALL sa_server_option ( 'Request_level_log_file',
> 'r.txt' >> ); CALL sa_server_option (
> 'Request_level_logging', >> 'SQL+hostvars' );
> >>
> >> -- Off...
> >> --CALL sa_server_option ( 'Request_level_logging',
> 'NONE' >> );
> >>
> >> Breck
> >>
> >>
> >> On 13 Feb 2004 09:12:18 -0800, modan wrote:
> >>
> >> >Hi, I'm going to be taking this up, as Dave is off on
> >> >holiday.
> >> >
> >> >Anyway, I have confirmed that the build is 1252.
> >> >
> >> >In answer to your other question, sa_conn_info shows
> one >> >blocked connection, not two.
> >> >
> >> >Both show ReqType of STMT_EXECUTE, and identical
> >> >lastreqtime, and UncmtOps = 0. The "unblocked" one
> shows >> a >LastIdle greater than the blocked one by
> either 1 or 2. >> >
> >> >Modan
> >> >
> >> >> So sa_conn_info showed two blocked connections,
> blocked >> on >> each other? That's a situation that
> should never >> actually >> be viewable IMHO.
> >> >>
> >> >> Are you sure you're on 1252, not build 1270?
> >> >>
> >> >> Please excuse my dumb questions, but I am chasing a
> >> >> related issue.
> >> >>
> >> >> Breck
> >> >>
> >> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
> >> >>
> >> >> >I found the "Deadlock Detected" worked when the
> >> threads >> were >in separate applications but not when
> in >> the same >> >application
> >> >> >
> >> >> >It did show a lock Cycle, Only 3 out of the 4
> Updates >> to >> the >database were processed by the
> server. >> >> >
> >> >> >Dave
> >> >> >
> >> >> >
> >> >> >> FWIW I just ran a deadlock test using two
> >> connections >> and >> got "Deadlock detected" as
> expected, >> in build 1252. >> >>
> >> >> >> What does sa_conn_info show when you run it on a
> 3rd >> >> >> connection? Does it show an actual lock cycle
> >> >> (deadlock)? >>
> >> >> >> Breck
> >> >> >>
> >> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
> >> >> >>
> >> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
> >> Build >> 1108 >> >
> >> >> >> >When Two Threads created a DeadLock, I get the
> >> Correct >> >> >Deadlock Error returned for one of the
> >> Threads, >> >> >
> >> >> >> >However when I install the Express Bug Fix,
> Update >> of >> >> 9.0.0 >to Build 1252
> >> >> >> >
> >> >> >> >I no longer get the DeadLock Error, The ASA
> >> Commands >> just >> >Hang.
> >> >> >> >
> >> >> >> >Has anyone else seen this ?
> >> >> >> >
> >> >> >> >Thanks in advance
> >> >> >> >
> >> >> >> >Dave
> >> >> >>
> >> >> >> --
> >> >> >> bcarter@risingroad.com
> >> >> >> Mobile and Distributed Enterprise Database
> >> Applications >> >> www.risingroad.com
> >> >>
> >> >> --
> >> >> bcarter@risingroad.com
> >> >> Mobile and Distributed Enterprise Database
> Applications >> >> www.risingroad.com
> >>
> >> --
> >> bcarter@risingroad.com
> >> Mobile and Distributed Enterprise Database Applications
> >> www.risingroad.com
>
> --
> SQL Anywhere Studio 9 Developer's Guide
> ISBN 1-55622-506-7
> bcarter@risingroad.com
> Mobile and Distributed Enterprise Database Applications
> www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-16 18:52:11.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <d332301k022klmp964n9sulogmavpnhsh4@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com><d0vp201gp4v7i28nld3e6sns4eirchcj0m@4ax.com> <402d0802.6961.846930886@sybase.com><0i8q209vqent7vl7jgo268kjiiv6u57m4e@4ax.com> <4030a8fd.2203.846930886@sybase.com><nfd1305ajpto3ab4tlj0pg66aqk7mkhp3c@4ax.com> <4030f93f.1f22.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: bcarter.sentex.ca
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
Date: 16 Feb 2004 10:52:11 -0800
X-Trace: forums-1-dub 1076957531 64.7.134.118 (16 Feb 2004 10:52:11 -0800)
X-Original-Trace: 16 Feb 2004 10:52:11 -0800, bcarter.sentex.ca
Lines: 316
X-Authenticated-User: TeamPS
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2518
Article PK: 5656

FWIW, the request level log will actually show *more* than the
transaction log in cases like this; for example, an UPDATE that is
blocked will show up in the request level log, but not in the
transaction log.

In this case, that is not important... I agree that the problem seems
to be on the client side of the client server boundary.

Breck

On 16 Feb 2004 08:58:05 -0800, modan wrote:

>As expected, translating the transaction log again shows no
>sign of the second thread's query actually making it as far
>as the DB engine, so it is looking increasingly likely that
>the bug is in the Adaptive Server Anywhere .NET Data
>Provider.
>
>> No, as far as I can tell only 2 of the updates finish, and
>> the third one is blocked (STMT_EXECUTE REQUEST but no
>> DONE).
>>
>> Also, I do not see any attempt by the first connection to
>> do its second update. Here is the trimmed-down request
>> level log, with comments...
>>
>> ===== Connection 1 completes UPDATE TableA...
>> 02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
>> 02/16 10:54:11.088 ** REQUEST conn: 304771280
>> STMT_PREPARE
>> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
>> 02/16 10:54:11.096 ** REQUEST conn: 304771280
>> STMT_PREPARE
>> "UPDATE TableA SET A = 2"
>> 02/16 10:54:11.096 ** DONE conn: 304771280
>> STMT_EXECUTE
>>
>> ===== Connection 1 completes UPDATE TableB, blocked on
>> UPDATE TableA...
>> 02/16 10:54:11.098 ** REQUEST conn: 304761224 CONNECT
>> 02/16 10:54:11.100 ** REQUEST conn: 304761224
>> STMT_PREPARE
>> "SET TEMPORARY OPTION ISOLATION_LEVEL=1"
>> 02/16 10:54:11.100 ** REQUEST conn: 304761224
>> STMT_PREPARE
>> "UPDATE TableB SET B = 2"
>> 02/16 10:54:11.101 ** DONE conn: 304761224
>> STMT_EXECUTE
>>
>> 02/16 10:54:11.102 ** REQUEST conn: 304761224
>> STMT_PREPARE
>> "UPDATE TableA SET A = 2"
>> 02/16 10:54:11.102 ** DONE conn: 304761224
>> STMT_PREPARE
>> Stmt=65540
>> 02/16 10:54:11.102 ** REQUEST conn: 304761224
>> STMT_EXECUTE
>> Stmt=-1
>>
>> Is there nothing happening past this point? If not, there
>> is no deadlock because the first connection (A.K.A.
>> 304771280) does not attempt to update TableB. What is
>> shown here is a simple block: the second connection
>> (A.K.A. 304761224) is waiting for the first connection to
>> release the lock(s) on TableA.
>>
>> Put it another way, you need to see 4 updates for there to
>> be a deadlock, and only 3 are shown.
>>
>> FWIW you can see more details of what is happening two
>> ways. First, CALL sa_locks() to see which connection has
>> locks on which table. Second, you can see the transaction
>> log in readable format by running this command (no need to
>> shut down the server first):
>>
>> "%ASANY9%\win32\dbtran.exe" -a -d -c "whatever" -s -y
>> whatever.sql
>>
>> The -f option is handy to reduce the amount of output,
>> especially if you do an explicit CHECKPOINT just before
>> starting your test run.
>>
>> Breck
>>
>>
>> On 16 Feb 2004 03:26:53 -0800, modan wrote:
>>
>> >Actually I am pretty sure there is a deadlock.
>> >
>> >The program creates two threads.
>> >
>> >First thread does
>> >
>> >update A followed by update B
>> >
>> >Second thread does
>> >
>> >update B followed by update A
>> >
>> >3 of the updates happen, according to the log, and both
>> >threads end up blocked (only one is blocked according to
>> >sa_conn_info though.
>> >
>> >It may be worth mentioning that this test app is written
>> in >C# and uses the .Net Data Provider.
>> >
>> >Request_level_logging is as follows:-
>> >
>> >02/16 10:54:11.085 ** REQUEST conn: 304771280 CONNECT
>> >02/16 10:54:11.085 ** CONNECT conn: 304771280 UID=DBA
>> >02/16 10:54:11.085 ** DONE conn: 304771280 CONNECT
>> > Conn=1744647885
>> >02/16 10:54:11.086 ** REQUEST conn: 304771280
>> > STMT_PREPARE "select db_property('charset')"
>> >02/16 10:54:11.086 ** DONE conn: 304771280
>> > STMT_PREPARE Stmt=65536
>> >02/16 10:54:11.086 ** REQUEST conn: 304771280
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.086 ** DONE conn: 304771280
>> STMT_EXECUTE >02/16 10:54:11.086 ** REQUEST conn:
>> > 304771280 STMT_DROP Stmt=65536
>> >02/16 10:54:11.086 ** DONE conn: 304771280 STMT_DROP
>> >02/16 10:54:11.086 ** REQUEST conn: 304771280
>> > STMT_PREPARE "select connection_property(
>> 'max_statement_count' >)"
>> >02/16 10:54:11.086 ** DONE conn: 304771280
>> > STMT_PREPARE Stmt=65537
>> >02/16 10:54:11.086 ** REQUEST conn: 304771280
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.086 ** DONE conn: 304771280
>> STMT_EXECUTE >02/16 10:54:11.088 ** REQUEST conn:
>> > 304771280 STMT_DROP Stmt=65537
>> >02/16 10:54:11.088 ** DONE conn: 304771280 STMT_DROP
>> >02/16 10:54:11.088 ** REQUEST conn: 304771280
>> > STMT_PREPARE "SET TEMPORARY OPTION
>> ISOLATION_LEVEL=1" >02/16 10:54:11.088 ** DONE conn:
>> > 304771280 STMT_PREPARE Stmt=65538
>> >02/16 10:54:11.088 ** REQUEST conn: 304771280
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.089 ** DONE conn: 304771280
>> STMT_EXECUTE >02/16 10:54:11.096 ** REQUEST conn:
>> > 304771280 STMT_DROP Stmt=65538
>> >02/16 10:54:11.096 ** DONE conn: 304771280 STMT_DROP
>> >02/16 10:54:11.096 ** REQUEST conn: 304771280
>> > STMT_PREPARE "UPDATE TableA SET A = 2"
>> >02/16 10:54:11.096 ** DONE conn: 304771280
>> > STMT_PREPARE Stmt=65539
>> >02/16 10:54:11.096 ** REQUEST conn: 304771280
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.096 ** DONE conn: 304771280
>> STMT_EXECUTE >02/16 10:54:11.098 ** REQUEST conn:
>> 304761224 CONNECT >02/16 10:54:11.098 ** CONNECT conn:
>> 304761224 UID=DBA >02/16 10:54:11.098 ** DONE conn:
>> > 304761224 CONNECT Conn=167015670
>> >02/16 10:54:11.098 ** REQUEST conn: 304761224
>> > STMT_PREPARE "select db_property('charset')"
>> >02/16 10:54:11.098 ** DONE conn: 304761224
>> > STMT_PREPARE Stmt=65536
>> >02/16 10:54:11.099 ** REQUEST conn: 304761224
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.099 ** DONE conn: 304761224
>> STMT_EXECUTE >02/16 10:54:11.099 ** REQUEST conn:
>> > 304761224 STMT_DROP Stmt=65536
>> >02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
>> >02/16 10:54:11.099 ** REQUEST conn: 304761224
>> > STMT_PREPARE "select connection_property(
>> 'max_statement_count' >)"
>> >02/16 10:54:11.099 ** DONE conn: 304761224
>> > STMT_PREPARE Stmt=65537
>> >02/16 10:54:11.099 ** REQUEST conn: 304761224
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.099 ** DONE conn: 304761224
>> STMT_EXECUTE >02/16 10:54:11.099 ** REQUEST conn:
>> > 304761224 STMT_DROP Stmt=65537
>> >02/16 10:54:11.099 ** DONE conn: 304761224 STMT_DROP
>> >02/16 10:54:11.100 ** REQUEST conn: 304761224
>> > STMT_PREPARE "SET TEMPORARY OPTION
>> ISOLATION_LEVEL=1" >02/16 10:54:11.100 ** DONE conn:
>> > 304761224 STMT_PREPARE Stmt=65538
>> >02/16 10:54:11.100 ** REQUEST conn: 304761224
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.100 ** DONE conn: 304761224
>> STMT_EXECUTE >02/16 10:54:11.100 ** REQUEST conn:
>> > 304761224 STMT_DROP Stmt=65538
>> >02/16 10:54:11.100 ** DONE conn: 304761224 STMT_DROP
>> >02/16 10:54:11.100 ** REQUEST conn: 304761224
>> > STMT_PREPARE "UPDATE TableB SET B = 2"
>> >02/16 10:54:11.101 ** DONE conn: 304761224
>> > STMT_PREPARE Stmt=65539
>> >02/16 10:54:11.101 ** REQUEST conn: 304761224
>> > STMT_EXECUTE Stmt=-1
>> >02/16 10:54:11.101 ** DONE conn: 304761224
>> STMT_EXECUTE >02/16 10:54:11.102 ** REQUEST conn:
>> > 304761224 STMT_DROP Stmt=65539
>> >02/16 10:54:11.102 ** DONE conn: 304761224 STMT_DROP
>> >02/16 10:54:11.102 ** REQUEST conn: 304761224
>> > STMT_PREPARE "UPDATE TableA SET A = 2"
>> >02/16 10:54:11.102 ** DONE conn: 304761224
>> > STMT_PREPARE Stmt=65540
>> >02/16 10:54:11.102 ** REQUEST conn: 304761224
>> > STMT_EXECUTE Stmt=-1
>> >
>> >> If there is only one blocked connection then it isn't a
>> >> deadlock.
>> >>
>> >> Which means either (a) there never was a deadlock or
>> (b) >> there was a deadlock and it got resolved. Can you
>> turn on >> request level logging to see what the sequence
>> of commands >> hitting the server are?
>> >>
>> >> -- On...
>> >> CALL sa_server_option ( 'Request_level_log_file',
>> 'r.txt' >> ); CALL sa_server_option (
>> 'Request_level_logging', >> 'SQL+hostvars' );
>> >>
>> >> -- Off...
>> >> --CALL sa_server_option ( 'Request_level_logging',
>> 'NONE' >> );
>> >>
>> >> Breck
>> >>
>> >>
>> >> On 13 Feb 2004 09:12:18 -0800, modan wrote:
>> >>
>> >> >Hi, I'm going to be taking this up, as Dave is off on
>> >> >holiday.
>> >> >
>> >> >Anyway, I have confirmed that the build is 1252.
>> >> >
>> >> >In answer to your other question, sa_conn_info shows
>> one >> >blocked connection, not two.
>> >> >
>> >> >Both show ReqType of STMT_EXECUTE, and identical
>> >> >lastreqtime, and UncmtOps = 0. The "unblocked" one
>> shows >> a >LastIdle greater than the blocked one by
>> either 1 or 2. >> >
>> >> >Modan
>> >> >
>> >> >> So sa_conn_info showed two blocked connections,
>> blocked >> on >> each other? That's a situation that
>> should never >> actually >> be viewable IMHO.
>> >> >>
>> >> >> Are you sure you're on 1252, not build 1270?
>> >> >>
>> >> >> Please excuse my dumb questions, but I am chasing a
>> >> >> related issue.
>> >> >>
>> >> >> Breck
>> >> >>
>> >> >> On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:
>> >> >>
>> >> >> >I found the "Deadlock Detected" worked when the
>> >> threads >> were >in separate applications but not when
>> in >> the same >> >application
>> >> >> >
>> >> >> >It did show a lock Cycle, Only 3 out of the 4
>> Updates >> to >> the >database were processed by the
>> server. >> >> >
>> >> >> >Dave
>> >> >> >
>> >> >> >
>> >> >> >> FWIW I just ran a deadlock test using two
>> >> connections >> and >> got "Deadlock detected" as
>> expected, >> in build 1252. >> >>
>> >> >> >> What does sa_conn_info show when you run it on a
>> 3rd >> >> >> connection? Does it show an actual lock cycle
>> >> >> (deadlock)? >>
>> >> >> >> Breck
>> >> >> >>
>> >> >> >> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>> >> >> >>
>> >> >> >> >Working with Sybase SQL Anywhere Verion 9.0.0.
>> >> Build >> 1108 >> >
>> >> >> >> >When Two Threads created a DeadLock, I get the
>> >> Correct >> >> >Deadlock Error returned for one of the
>> >> Threads, >> >> >
>> >> >> >> >However when I install the Express Bug Fix,
>> Update >> of >> >> 9.0.0 >to Build 1252
>> >> >> >> >
>> >> >> >> >I no longer get the DeadLock Error, The ASA
>> >> Commands >> just >> >Hang.
>> >> >> >> >
>> >> >> >> >Has anyone else seen this ?
>> >> >> >> >
>> >> >> >> >Thanks in advance
>> >> >> >> >
>> >> >> >> >Dave
>> >> >> >>
>> >> >> >> --
>> >> >> >> bcarter@risingroad.com
>> >> >> >> Mobile and Distributed Enterprise Database
>> >> Applications >> >> www.risingroad.com
>> >> >>
>> >> >> --
>> >> >> bcarter@risingroad.com
>> >> >> Mobile and Distributed Enterprise Database
>> Applications >> >> www.risingroad.com
>> >>
>> >> --
>> >> bcarter@risingroad.com
>> >> Mobile and Distributed Enterprise Database Applications
>> >> www.risingroad.com
>>
>> --
>> SQL Anywhere Studio 9 Developer's Guide
>> ISBN 1-55622-506-7
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
SQL Anywhere Studio 9 Developer's Guide
Buy the book: http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


Breck Carter [TeamSybase] Posted on 2004-02-13 17:09:29.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <p10q20p1iakb970lged4s19cpt6igmscmb@4ax.com>
References: <402ca611.6481.846930886@sybase.com><dljp20ptapfq3l6d74fej5ovhmpvirbm5d@4ax.com> <402cfaba.69da.846930886@sybase.com>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
X-Original-Trace: 13 Feb 2004 09:20:32 -0800, bcarter.sentex.ca
Lines: 53
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 13 Feb 2004 08:51:19 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 13 Feb 2004 09:09:29 -0800
X-Trace: forums-1-dub 1076692169 10.22.108.75 (13 Feb 2004 09:09:29 -0800)
X-Original-Trace: 13 Feb 2004 09:09:29 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2492
Article PK: 5630

Forget the question about the build number, but please confirm that
sa_conn_info actually showed two connections blocked by each other.

Breck

On 13 Feb 2004 08:26:34 -0800, Dave Savage wrote:

>I found the "Deadlock Detected" worked when the threads were
>in separate applications but not when in the same
>application
>
>It did show a lock Cycle, Only 3 out of the 4 Updates to the
>database were processed by the server.
>
>Dave
>
>
>> FWIW I just ran a deadlock test using two connections and
>> got "Deadlock detected" as expected, in build 1252.
>>
>> What does sa_conn_info show when you run it on a 3rd
>> connection? Does it show an actual lock cycle (deadlock)?
>>
>> Breck
>>
>> On 13 Feb 2004 02:25:21 -0800, Dave Savage wrote:
>>
>> >Working with Sybase SQL Anywhere Verion 9.0.0. Build 1108
>> >
>> >When Two Threads created a DeadLock, I get the Correct
>> >Deadlock Error returned for one of the Threads,
>> >
>> >However when I install the Express Bug Fix, Update of
>> 9.0.0 >to Build 1252
>> >
>> >I no longer get the DeadLock Error, The ASA Commands just
>> >Hang.
>> >
>> >Has anyone else seen this ?
>> >
>> >Thanks in advance
>> >
>> >Dave
>>
>> --
>> bcarter@risingroad.com
>> Mobile and Distributed Enterprise Database Applications
>> www.risingroad.com

--
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com


Stephen Rice Posted on 2004-02-18 16:15:23.0Z
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
From: Stephen Rice <srice_nospam@ianywhere.com>
References: <402cfaba.69da.846930886@sybase.com> <p10q20p1iakb970lged4s19cpt6igmscmb@4ax.com>
Organization: iAnywhere Solutions
Message-ID: <Xns94936F58391F3sricenospamianywhere@10.22.241.106>
User-Agent: Xnews/5.04.25
X-Original-NNTP-Posting-Host: srice-pc.sybase.com
X-Original-Trace: 18 Feb 2004 08:26:45 -0800, srice-pc.sybase.com
Lines: 19
X-Original-NNTP-Posting-Host: forums-2-dub.sybase.com
X-Original-Trace: 18 Feb 2004 07:56:45 -0800, forums-2-dub.sybase.com
NNTP-Posting-Host: forums-master.sybase.com
X-Original-NNTP-Posting-Host: forums-master.sybase.com
Date: 18 Feb 2004 08:15:23 -0800
X-Trace: forums-1-dub 1077120923 10.22.108.75 (18 Feb 2004 08:15:23 -0800)
X-Original-Trace: 18 Feb 2004 08:15:23 -0800, forums-master.sybase.com
X-Authenticated-User: ngsysop
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2542
Article PK: 5681

I'm jumping in late but I have an "obvious" question that I don't see asked
or answered (which means it's not really obvious)....

Is each thread using its own connection? I'm not a .NET expert but I do
know with our other APIs each thread needs its own connection if they are
going to try and execute SQL statements simultaneously.

/steve

--
Stephen Rice
Technical Services Manager
iAnywhere Solutions

- Please Post-
Whitepapers, TechDocs, and bug fixes are all available through the
iAnywhere
Developer Community at www.ianywhere.com/developer


Breck Carter [TeamSybase] Posted on 2004-02-18 17:00:50.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__bcarter@risingroad.com>
Newsgroups: ianywhere.public.general
Subject: Re: No Deadlock Error reported with EBF Update of 9.0.0 to Build 1252
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__bcarter@risingroad.com
Message-ID: <3d5730137r6qv0glld7gi35ns7u7ckm1mc@4ax.com>
References: <402cfaba.69da.846930886@sybase.com> <p10q20p1iakb970lged4s19cpt6igmscmb@4ax.com> <Xns94936F58391F3sricenospamianywhere@10.22.241.106>
X-Newsreader: Forte Agent 1.8/32.548
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: bcarter.sentex.ca
X-Original-NNTP-Posting-Host: bcarter.sentex.ca
Date: 18 Feb 2004 09:00:50 -0800
X-Trace: forums-1-dub 1077123650 64.7.134.118 (18 Feb 2004 09:00:50 -0800)
X-Original-Trace: 18 Feb 2004 09:00:50 -0800, bcarter.sentex.ca
Lines: 20
X-Authenticated-User: TeamPS
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub ianywhere.public.general:2545
Article PK: 5683

The request level log posted earlier shows 2 connections.

On 18 Feb 2004 08:15:23 -0800, Stephen Rice

<srice_nospam@ianywhere.com> wrote:

>I'm jumping in late but I have an "obvious" question that I don't see asked
>or answered (which means it's not really obvious)....
>
>Is each thread using its own connection? I'm not a .NET expert but I do
>know with our other APIs each thread needs its own connection if they are
>going to try and execute SQL statements simultaneously.
>
>/steve

--
SQL Anywhere Studio 9 Developer's Guide
Buy the book: http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
bcarter@risingroad.com
Mobile and Distributed Enterprise Database Applications
www.risingroad.com