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.

Poxy table connection terminated?!

19 posts in General Discussion (old) Last posting was on 2008-10-30 08:26:18.0Z
Markus KARG Posted on 2008-10-07 07:11:36.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
Subject: Poxy table connection terminated?!
Lines: 22
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48eb0ba8$1@forums-1-dub>
Date: 7 Oct 2008 00:11:36 -0700
X-Trace: forums-1-dub 1223363496 10.22.241.152 (7 Oct 2008 00:11:36 -0700)
X-Original-Trace: 7 Oct 2008 00:11:36 -0700, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:179
Article PK: 866624

We are using GlassFishv2 Application Server to connect to a SQL Any 11
database (using iAnywhere JDBC driver). This connection works well. We see
all data from data database. That database has proxy tables hosted by a
second SQL Any 11 database on the same database server engine. GlassFish can
see that data, too (obivously).

We do not query any data over the night. At the next morning, we still can
see all data from the database, but the data from the proxy tables does not
occur. Instead the JDBC Driver throws a SQL Exception telling is that the
foreign server cannot be reached: "Connection terminated" it says.

When we go to Sybase Central we can see the data in the proxy tables! Only
the JDBC driver is not seeting that data. VERY STRANGE!

What can we do? Is that a bug?

Please HELP!

Thanks
Markus


Breck Carter [TeamSybase] Posted on 2008-10-07 12:29:25.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com>
Newsgroups: sybase.public.sqlanywhere
Subject: Re: Poxy table connection terminated?!
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__breck.carter@gmail.com
Message-ID: <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com>
References: <48eb0ba8$1@forums-1-dub>
X-Newsreader: Forte Agent 2.0/32.640
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Date: 7 Oct 2008 05:29:25 -0700
X-Trace: forums-1-dub 1223382565 10.22.241.152 (7 Oct 2008 05:29:25 -0700)
X-Original-Trace: 7 Oct 2008 05:29:25 -0700, vip152.sybase.com
Lines: 40
X-Authenticated-User: TeamSybase
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:180
Article PK: 866626

It might be an idle timeout issue on the proxy connection between the
two servers which is probably using ODBC; see this connection
parameter which you can put in the ODBC DSN or in the CREATE SERVER
command:

SQL Anywhere Server - Database Administration » Starting and
Connecting to Your Database » Connection parameters and network
protocol options » Connection parameters » Idle connection parameter

On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:

>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>database (using iAnywhere JDBC driver). This connection works well. We see
>all data from data database. That database has proxy tables hosted by a
>second SQL Any 11 database on the same database server engine. GlassFish can
>see that data, too (obivously).
>
>We do not query any data over the night. At the next morning, we still can
>see all data from the database, but the data from the proxy tables does not
>occur. Instead the JDBC Driver throws a SQL Exception telling is that the
>foreign server cannot be reached: "Connection terminated" it says.
>
>When we go to Sybase Central we can see the data in the proxy tables! Only
>the JDBC driver is not seeting that data. VERY STRANGE!
>
>What can we do? Is that a bug?
>
>Please HELP!
>
>Thanks
>Markus
>

--
Breck Carter http://sqlanywhere.blogspot.com/

RisingRoad SQL Anywhere and MobiLink Professional Services
breck.carter@risingroad.com


Markus KARG Posted on 2008-10-08 06:24:01.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com>
Subject: Re: Poxy table connection terminated?!
Lines: 68
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48ec5201$1@forums-1-dub>
Date: 7 Oct 2008 23:24:01 -0700
X-Trace: forums-1-dub 1223447041 10.22.241.152 (7 Oct 2008 23:24:01 -0700)
X-Original-Trace: 7 Oct 2008 23:24:01 -0700, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:181
Article PK: 866627

Breck,

it is THE SAME server (same instance of dbsrv11.exe on the same host), but
just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults, we
did not change them (no options set, neither in dbsrv11.exe nor in the DSN)!
Also, why is it working in Sybase Central but not in JDBC (the proxy should
be transparent to the client's connection, obviously). Also, the default is
240 minutes, but we encountered the problem after just 60 minutes. Maybe a
bug? BTW, we had similar (but less often) problems back in SQL Anywhere 8
and JOnAS Application Server, so it is not a problem of the Application
Server as it seems. And: This problems occurs at a lot of our customers
LANs, not only in our lab. For me, it looks like a bug. I mean, when using
defaults then a proxy connection on the same host should never die,
obviously, or at least it should automatically reconnect... And no, the
number of connections is just a dozen, neither RAM nor CPU nor HDD is
utilizing more than 50% of their possible load.

Anything you want me to check?

Thanks!
Markus

"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...

> It might be an idle timeout issue on the proxy connection between the
> two servers which is probably using ODBC; see this connection
> parameter which you can put in the ODBC DSN or in the CREATE SERVER
> command:
>
> SQL Anywhere Server - Database Administration » Starting and
> Connecting to Your Database » Connection parameters and network
> protocol options » Connection parameters » Idle connection parameter
>
>
>
> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>
>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>database (using iAnywhere JDBC driver). This connection works well. We see
>>all data from data database. That database has proxy tables hosted by a
>>second SQL Any 11 database on the same database server engine. GlassFish
>>can
>>see that data, too (obivously).
>>
>>We do not query any data over the night. At the next morning, we still can
>>see all data from the database, but the data from the proxy tables does
>>not
>>occur. Instead the JDBC Driver throws a SQL Exception telling is that the
>>foreign server cannot be reached: "Connection terminated" it says.
>>
>>When we go to Sybase Central we can see the data in the proxy tables! Only
>>the JDBC driver is not seeting that data. VERY STRANGE!
>>
>>What can we do? Is that a bug?
>>
>>Please HELP!
>>
>>Thanks
>>Markus
>>
>
> --
> Breck Carter http://sqlanywhere.blogspot.com/
>
> RisingRoad SQL Anywhere and MobiLink Professional Services
> breck.carter@risingroad.com


Breck Carter [TeamSybase] Posted on 2008-10-08 12:10:10.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com>
Newsgroups: sybase.public.sqlanywhere
Subject: Re: Poxy table connection terminated?!
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__breck.carter@gmail.com
Message-ID: <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com>
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub>
X-Newsreader: Forte Agent 2.0/32.640
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Date: 8 Oct 2008 05:10:10 -0700
X-Trace: forums-1-dub 1223467810 10.22.241.152 (8 Oct 2008 05:10:10 -0700)
X-Original-Trace: 8 Oct 2008 05:10:10 -0700, vip152.sybase.com
Lines: 81
X-Authenticated-User: TeamSybase
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:182
Article PK: 866628

Please show us the exact error message.

Is it the ODBC proxy connection from one database to the other that is
being disconnected, or the "primary" JDBC connection from the outside
world into the first database?

On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:

>Breck,
>
>it is THE SAME server (same instance of dbsrv11.exe on the same host), but
>just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults, we
>did not change them (no options set, neither in dbsrv11.exe nor in the DSN)!
>Also, why is it working in Sybase Central but not in JDBC (the proxy should
>be transparent to the client's connection, obviously). Also, the default is
>240 minutes, but we encountered the problem after just 60 minutes. Maybe a
>bug? BTW, we had similar (but less often) problems back in SQL Anywhere 8
>and JOnAS Application Server, so it is not a problem of the Application
>Server as it seems. And: This problems occurs at a lot of our customers
>LANs, not only in our lab. For me, it looks like a bug. I mean, when using
>defaults then a proxy connection on the same host should never die,
>obviously, or at least it should automatically reconnect... And no, the
>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>utilizing more than 50% of their possible load.
>
>Anything you want me to check?
>
>Thanks!
>Markus
>
>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>> It might be an idle timeout issue on the proxy connection between the
>> two servers which is probably using ODBC; see this connection
>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>> command:
>>
>> SQL Anywhere Server - Database Administration » Starting and
>> Connecting to Your Database » Connection parameters and network
>> protocol options » Connection parameters » Idle connection parameter
>>
>>
>>
>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>
>>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>>database (using iAnywhere JDBC driver). This connection works well. We see
>>>all data from data database. That database has proxy tables hosted by a
>>>second SQL Any 11 database on the same database server engine. GlassFish
>>>can
>>>see that data, too (obivously).
>>>
>>>We do not query any data over the night. At the next morning, we still can
>>>see all data from the database, but the data from the proxy tables does
>>>not
>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that the
>>>foreign server cannot be reached: "Connection terminated" it says.
>>>
>>>When we go to Sybase Central we can see the data in the proxy tables! Only
>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>
>>>What can we do? Is that a bug?
>>>
>>>Please HELP!
>>>
>>>Thanks
>>>Markus
>>>
>>
>> --
>> Breck Carter http://sqlanywhere.blogspot.com/
>>
>> RisingRoad SQL Anywhere and MobiLink Professional Services
>> breck.carter@risingroad.com
>

--
Breck Carter http://sqlanywhere.blogspot.com/

RisingRoad SQL Anywhere and MobiLink Professional Services
breck.carter@risingroad.com


Markus KARG Posted on 2008-10-09 14:01:23.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com>
Subject: Re: Poxy table connection terminated?!
Lines: 118
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Original
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <48ee0eb3$1@forums-1-dub>
Date: 9 Oct 2008 07:01:23 -0700
X-Trace: forums-1-dub 1223560883 10.22.241.152 (9 Oct 2008 07:01:23 -0700)
X-Original-Trace: 9 Oct 2008 07:01:23 -0700, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:186
Article PK: 866629

The connection from JDBC client software to "base" database is still working
(I can run SELECTs on its BASE tables), but the ODBC proxy connection from
"base" database to proxied database is broken (running SELECTs on PROXY
tables throw exceptions)! That is a bit weird, since it is the same
dbsrv11.exe instance, so either it can be connected or it cannot. For me it
seems as if the "proxy connection" has its own idle timer -- that runs into
a timeout because my last accesses was on BASE but not on PROXY tables. If
that is true, that would be strange, because virtually everybody would
expect that the proxy connection is kept alive as long as the "base"
connection is functional.

The message is:

java.sql.SQLException: [Sybase][ODBC Driver][SQL Anywhere]Server 'FMEA':
[Sybase][ODBC Driver][SQL Anywhere]Verbindung wurde beendet
Error Code: -660

("Verbindung wurde beendet" means "Connection was terminated")

BTW, it is not dependend of our client application, or server or LAN -- our
customers do have this behaviour, too...

Thanks!
Markus

"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...

> Please show us the exact error message.
>
> Is it the ODBC proxy connection from one database to the other that is
> being disconnected, or the "primary" JDBC connection from the outside
> world into the first database?
>
> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>
>>Breck,
>>
>>it is THE SAME server (same instance of dbsrv11.exe on the same host), but
>>just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults,
>>we
>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>DSN)!
>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>should
>>be transparent to the client's connection, obviously). Also, the default
>>is
>>240 minutes, but we encountered the problem after just 60 minutes. Maybe a
>>bug? BTW, we had similar (but less often) problems back in SQL Anywhere 8
>>and JOnAS Application Server, so it is not a problem of the Application
>>Server as it seems. And: This problems occurs at a lot of our customers
>>LANs, not only in our lab. For me, it looks like a bug. I mean, when using
>>defaults then a proxy connection on the same host should never die,
>>obviously, or at least it should automatically reconnect... And no, the
>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>utilizing more than 50% of their possible load.
>>
>>Anything you want me to check?
>>
>>Thanks!
>>Markus
>>
>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>> It might be an idle timeout issue on the proxy connection between the
>>> two servers which is probably using ODBC; see this connection
>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>> command:
>>>
>>> SQL Anywhere Server - Database Administration » Starting and
>>> Connecting to Your Database » Connection parameters and network
>>> protocol options » Connection parameters » Idle connection parameter
>>>
>>>
>>>
>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>
>>>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>>>database (using iAnywhere JDBC driver). This connection works well. We
>>>>see
>>>>all data from data database. That database has proxy tables hosted by a
>>>>second SQL Any 11 database on the same database server engine. GlassFish
>>>>can
>>>>see that data, too (obivously).
>>>>
>>>>We do not query any data over the night. At the next morning, we still
>>>>can
>>>>see all data from the database, but the data from the proxy tables does
>>>>not
>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that
>>>>the
>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>
>>>>When we go to Sybase Central we can see the data in the proxy tables!
>>>>Only
>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>
>>>>What can we do? Is that a bug?
>>>>
>>>>Please HELP!
>>>>
>>>>Thanks
>>>>Markus
>>>>
>>>
>>> --
>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>
>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>> breck.carter@risingroad.com
>>
>
> --
> Breck Carter http://sqlanywhere.blogspot.com/
>
> RisingRoad SQL Anywhere and MobiLink Professional Services
> breck.carter@risingroad.com


Markus KARG Posted on 2008-10-27 09:47:29.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com>
Subject: Re: Poxy table connection terminated?!
Lines: 138
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Original
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <49058e31$1@forums-1-dub>
Date: 27 Oct 2008 01:47:29 -0800
X-Trace: forums-1-dub 1225100849 10.22.241.152 (27 Oct 2008 01:47:29 -0800)
X-Original-Trace: 27 Oct 2008 01:47:29 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:203
Article PK: 866647

Dear SQL Anywhere Professionals,

I am in contact with Guido Farrenberg from Sybase Germany and we do not see
a solution for the problem, still. The problem is:

* Application opens JDBC connection to dbsrv11 server (in fact,
"application" is GlassFish Application Server provided by Sun and Oracle).
* dbsrv11 server in turn opens connection to itself for accessing the proxy
table hosted in another database on the same dbsrv11 instance.
* Both BASE and PROXY tables get used a lot without problems.
* Then, over night, when nobody works with the application, no use of BASE
or PROXY tables occurs.
* Next morning, at first use, application does SELECT * FROM MyProxyTable to
detect whether the pooled connection sill works and gets NO error message
from the JDBC driver (Guido assumes that the PrepareStatement command is
executed correctly even on a broken connection).
* Then, reading data from BASE tables works well -- the connection to the
"front end" server is working well.
* But reading data from PROXY tables crash -- the connection from the "front
end" server to the "back end" answering the proxy requests is down.

WHY???

Guido says there is no solution.

I must set idle timeout to zero for the "internal" connection. Certainly
that prevents the loss of the "internal" connection from "front end" server
to "back end" server, but it leads to a lot of side effects -- starting from
lots of "zombie" connections kept til doomsday, to potentially never removed
locks in case of crashed connections etc.

So what we are seeking for desparately (!!!) is a *real* solution. From the
user's view, this is clearly a bug in dbsrv11: As long as BASE tables can be
read, also PROXY tables must be open for reading over THE SAME CONNECTION --
the "internal" connection must not timeout before the "original" connection
is running into timeout, obviously. No application will ever understand why
a SELECT on the BASE table is working, while a PROXY table "suddenly"
returns "Connection lost" at the same time, OVER THE SAME CONNECTION.

Please, can you provide a solution for that? Our customers are near to kill
me. :-( What to do???

Thanks
Markus

"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...

> Please show us the exact error message.
>
> Is it the ODBC proxy connection from one database to the other that is
> being disconnected, or the "primary" JDBC connection from the outside
> world into the first database?
>
> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>
>>Breck,
>>
>>it is THE SAME server (same instance of dbsrv11.exe on the same host), but
>>just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults,
>>we
>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>DSN)!
>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>should
>>be transparent to the client's connection, obviously). Also, the default
>>is
>>240 minutes, but we encountered the problem after just 60 minutes. Maybe a
>>bug? BTW, we had similar (but less often) problems back in SQL Anywhere 8
>>and JOnAS Application Server, so it is not a problem of the Application
>>Server as it seems. And: This problems occurs at a lot of our customers
>>LANs, not only in our lab. For me, it looks like a bug. I mean, when using
>>defaults then a proxy connection on the same host should never die,
>>obviously, or at least it should automatically reconnect... And no, the
>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>utilizing more than 50% of their possible load.
>>
>>Anything you want me to check?
>>
>>Thanks!
>>Markus
>>
>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>> It might be an idle timeout issue on the proxy connection between the
>>> two servers which is probably using ODBC; see this connection
>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>> command:
>>>
>>> SQL Anywhere Server - Database Administration » Starting and
>>> Connecting to Your Database » Connection parameters and network
>>> protocol options » Connection parameters » Idle connection parameter
>>>
>>>
>>>
>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>
>>>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>>>database (using iAnywhere JDBC driver). This connection works well. We
>>>>see
>>>>all data from data database. That database has proxy tables hosted by a
>>>>second SQL Any 11 database on the same database server engine. GlassFish
>>>>can
>>>>see that data, too (obivously).
>>>>
>>>>We do not query any data over the night. At the next morning, we still
>>>>can
>>>>see all data from the database, but the data from the proxy tables does
>>>>not
>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that
>>>>the
>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>
>>>>When we go to Sybase Central we can see the data in the proxy tables!
>>>>Only
>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>
>>>>What can we do? Is that a bug?
>>>>
>>>>Please HELP!
>>>>
>>>>Thanks
>>>>Markus
>>>>
>>>
>>> --
>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>
>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>> breck.carter@risingroad.com
>>
>
> --
> Breck Carter http://sqlanywhere.blogspot.com/
>
> RisingRoad SQL Anywhere and MobiLink Professional Services
> breck.carter@risingroad.com


Jason Hinsperger Posted on 2008-10-27 14:49:45.0Z
From: "Jason Hinsperger" <NO_jason_hinsperger_SPAM@hotmail.com>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 163
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4905d509$1@forums-1-dub>
Date: 27 Oct 2008 06:49:45 -0800
X-Trace: forums-1-dub 1225118985 10.22.241.152 (27 Oct 2008 06:49:45 -0800)
X-Original-Trace: 27 Oct 2008 06:49:45 -0800, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:205
Article PK: 866648

If setting the connection timeout to 0 (never) is working for you, have you
considered/tried setting it to a more realistic value?
If the connections are timing out overnight, perhaps increasing the timeout
value to 12 hours would be enough to resolve the issue in most cases and
avoid the other problems of having no timeout set.

--
Jason Hinsperger
Product Manager
Sybase iAnywhere
http://iablog.sybase.com/hinsperg

SQL Anywhere Developer Community
http://www.ianywhere.com/developer

"Markus KARG" <karg@quipsy.de> wrote in message
news:49058e31$1@forums-1-dub...
> Dear SQL Anywhere Professionals,
>
> I am in contact with Guido Farrenberg from Sybase Germany and we do not
> see a solution for the problem, still. The problem is:
>
> * Application opens JDBC connection to dbsrv11 server (in fact,
> "application" is GlassFish Application Server provided by Sun and Oracle).
> * dbsrv11 server in turn opens connection to itself for accessing the
> proxy table hosted in another database on the same dbsrv11 instance.
> * Both BASE and PROXY tables get used a lot without problems.
> * Then, over night, when nobody works with the application, no use of BASE
> or PROXY tables occurs.
> * Next morning, at first use, application does SELECT * FROM MyProxyTable
> to detect whether the pooled connection sill works and gets NO error
> message from the JDBC driver (Guido assumes that the PrepareStatement
> command is executed correctly even on a broken connection).
> * Then, reading data from BASE tables works well -- the connection to the
> "front end" server is working well.
> * But reading data from PROXY tables crash -- the connection from the
> "front end" server to the "back end" answering the proxy requests is down.
>
> WHY???
>
> Guido says there is no solution.
>
> I must set idle timeout to zero for the "internal" connection. Certainly
> that prevents the loss of the "internal" connection from "front end"
> server to "back end" server, but it leads to a lot of side effects --
> starting from lots of "zombie" connections kept til doomsday, to
> potentially never removed locks in case of crashed connections etc.
>
> So what we are seeking for desparately (!!!) is a *real* solution. From
> the user's view, this is clearly a bug in dbsrv11: As long as BASE tables
> can be read, also PROXY tables must be open for reading over THE SAME
> CONNECTION -- the "internal" connection must not timeout before the
> "original" connection is running into timeout, obviously. No application
> will ever understand why a SELECT on the BASE table is working, while a
> PROXY table "suddenly" returns "Connection lost" at the same time, OVER
> THE SAME CONNECTION.
>
> Please, can you provide a solution for that? Our customers are near to
> kill me. :-( What to do???
>
> Thanks
> Markus
>
> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
> Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...
>> Please show us the exact error message.
>>
>> Is it the ODBC proxy connection from one database to the other that is
>> being disconnected, or the "primary" JDBC connection from the outside
>> world into the first database?
>>
>> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>
>>>Breck,
>>>
>>>it is THE SAME server (same instance of dbsrv11.exe on the same host),
>>>but
>>>just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults,
>>>we
>>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>>DSN)!
>>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>>should
>>>be transparent to the client's connection, obviously). Also, the default
>>>is
>>>240 minutes, but we encountered the problem after just 60 minutes. Maybe
>>>a
>>>bug? BTW, we had similar (but less often) problems back in SQL Anywhere 8
>>>and JOnAS Application Server, so it is not a problem of the Application
>>>Server as it seems. And: This problems occurs at a lot of our customers
>>>LANs, not only in our lab. For me, it looks like a bug. I mean, when
>>>using
>>>defaults then a proxy connection on the same host should never die,
>>>obviously, or at least it should automatically reconnect... And no, the
>>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>>utilizing more than 50% of their possible load.
>>>
>>>Anything you want me to check?
>>>
>>>Thanks!
>>>Markus
>>>
>>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>>> It might be an idle timeout issue on the proxy connection between the
>>>> two servers which is probably using ODBC; see this connection
>>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>>> command:
>>>>
>>>> SQL Anywhere Server - Database Administration » Starting and
>>>> Connecting to Your Database » Connection parameters and network
>>>> protocol options » Connection parameters » Idle connection parameter
>>>>
>>>>
>>>>
>>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>
>>>>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>>>>database (using iAnywhere JDBC driver). This connection works well. We
>>>>>see
>>>>>all data from data database. That database has proxy tables hosted by a
>>>>>second SQL Any 11 database on the same database server engine.
>>>>>GlassFish
>>>>>can
>>>>>see that data, too (obivously).
>>>>>
>>>>>We do not query any data over the night. At the next morning, we still
>>>>>can
>>>>>see all data from the database, but the data from the proxy tables does
>>>>>not
>>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that
>>>>>the
>>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>>
>>>>>When we go to Sybase Central we can see the data in the proxy tables!
>>>>>Only
>>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>>
>>>>>What can we do? Is that a bug?
>>>>>
>>>>>Please HELP!
>>>>>
>>>>>Thanks
>>>>>Markus
>>>>>
>>>>
>>>> --
>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>
>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>> breck.carter@risingroad.com
>>>
>>
>> --
>> Breck Carter http://sqlanywhere.blogspot.com/
>>
>> RisingRoad SQL Anywhere and MobiLink Professional Services
>> breck.carter@risingroad.com
>
>


Markus KARG Posted on 2008-10-27 20:02:45.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub> <4905d509$1@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 196
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <49061e65@forums-1-dub>
Date: 27 Oct 2008 12:02:45 -0800
X-Trace: forums-1-dub 1225137765 10.22.241.152 (27 Oct 2008 12:02:45 -0800)
X-Original-Trace: 27 Oct 2008 12:02:45 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:207
Article PK: 866651

Jason,

thanks for your kind help. :-)

> If setting the connection timeout to 0 (never) is working for you, have
> you considered/tried setting it to a more realistic value?

Well, what is a realistic value? Since the client is an application server
pooling connections, we just don't know how long the connection is used or
unused. The application server in fact is configured to close any unused
connections after five minutes, so we are upset that the BASE tables can be
read without any problems, while the PROXY tables cannot.

> If the connections are timing out overnight, perhaps increasing the
> timeout value to 12 hours would be enough to resolve the issue in most
> cases and avoid the other problems of having no timeout set.

Guide did not tell us to do so, so we abstained from changing his decision.
But your tip sounds good. Maybe we will try to do so. My fear is that
changing the value from 240 minutes to 12 hours might just shift the problem
into the future instead of completely solving it.

What we are interested in is a clear explanation what the cause is, and a
100% correctly working solution. Still nobody could tell us why BASE tables
can be read, but PROXY tables produce connection timeout. As I said, it is a
connection from and to the same instance of dbsrv11. So either both time
out, or none -- or the proxy should be reconnect automatically.

Regards
Markus

>
> --
> Jason Hinsperger
> Product Manager
> Sybase iAnywhere
> http://iablog.sybase.com/hinsperg
>
> SQL Anywhere Developer Community
> http://www.ianywhere.com/developer
>
>
> "Markus KARG" <karg@quipsy.de> wrote in message
> news:49058e31$1@forums-1-dub...
>> Dear SQL Anywhere Professionals,
>>
>> I am in contact with Guido Farrenberg from Sybase Germany and we do not
>> see a solution for the problem, still. The problem is:
>>
>> * Application opens JDBC connection to dbsrv11 server (in fact,
>> "application" is GlassFish Application Server provided by Sun and
>> Oracle).
>> * dbsrv11 server in turn opens connection to itself for accessing the
>> proxy table hosted in another database on the same dbsrv11 instance.
>> * Both BASE and PROXY tables get used a lot without problems.
>> * Then, over night, when nobody works with the application, no use of
>> BASE or PROXY tables occurs.
>> * Next morning, at first use, application does SELECT * FROM MyProxyTable
>> to detect whether the pooled connection sill works and gets NO error
>> message from the JDBC driver (Guido assumes that the PrepareStatement
>> command is executed correctly even on a broken connection).
>> * Then, reading data from BASE tables works well -- the connection to the
>> "front end" server is working well.
>> * But reading data from PROXY tables crash -- the connection from the
>> "front end" server to the "back end" answering the proxy requests is
>> down.
>>
>> WHY???
>>
>> Guido says there is no solution.
>>
>> I must set idle timeout to zero for the "internal" connection. Certainly
>> that prevents the loss of the "internal" connection from "front end"
>> server to "back end" server, but it leads to a lot of side effects --
>> starting from lots of "zombie" connections kept til doomsday, to
>> potentially never removed locks in case of crashed connections etc.
>>
>> So what we are seeking for desparately (!!!) is a *real* solution. From
>> the user's view, this is clearly a bug in dbsrv11: As long as BASE tables
>> can be read, also PROXY tables must be open for reading over THE SAME
>> CONNECTION -- the "internal" connection must not timeout before the
>> "original" connection is running into timeout, obviously. No application
>> will ever understand why a SELECT on the BASE table is working, while a
>> PROXY table "suddenly" returns "Connection lost" at the same time, OVER
>> THE SAME CONNECTION.
>>
>> Please, can you provide a solution for that? Our customers are near to
>> kill me. :-( What to do???
>>
>> Thanks
>> Markus
>>
>> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>> Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...
>>> Please show us the exact error message.
>>>
>>> Is it the ODBC proxy connection from one database to the other that is
>>> being disconnected, or the "primary" JDBC connection from the outside
>>> world into the first database?
>>>
>>> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>
>>>>Breck,
>>>>
>>>>it is THE SAME server (same instance of dbsrv11.exe on the same host),
>>>>but
>>>>just two databases! The ODBC DSN and the dbsrv11.exe uses only defaults,
>>>>we
>>>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>>>DSN)!
>>>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>>>should
>>>>be transparent to the client's connection, obviously). Also, the default
>>>>is
>>>>240 minutes, but we encountered the problem after just 60 minutes. Maybe
>>>>a
>>>>bug? BTW, we had similar (but less often) problems back in SQL Anywhere
>>>>8
>>>>and JOnAS Application Server, so it is not a problem of the Application
>>>>Server as it seems. And: This problems occurs at a lot of our customers
>>>>LANs, not only in our lab. For me, it looks like a bug. I mean, when
>>>>using
>>>>defaults then a proxy connection on the same host should never die,
>>>>obviously, or at least it should automatically reconnect... And no, the
>>>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>>>utilizing more than 50% of their possible load.
>>>>
>>>>Anything you want me to check?
>>>>
>>>>Thanks!
>>>>Markus
>>>>
>>>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>>>> It might be an idle timeout issue on the proxy connection between the
>>>>> two servers which is probably using ODBC; see this connection
>>>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>>>> command:
>>>>>
>>>>> SQL Anywhere Server - Database Administration » Starting and
>>>>> Connecting to Your Database » Connection parameters and network
>>>>> protocol options » Connection parameters » Idle connection parameter
>>>>>
>>>>>
>>>>>
>>>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>>
>>>>>>We are using GlassFishv2 Application Server to connect to a SQL Any 11
>>>>>>database (using iAnywhere JDBC driver). This connection works well. We
>>>>>>see
>>>>>>all data from data database. That database has proxy tables hosted by
>>>>>>a
>>>>>>second SQL Any 11 database on the same database server engine.
>>>>>>GlassFish
>>>>>>can
>>>>>>see that data, too (obivously).
>>>>>>
>>>>>>We do not query any data over the night. At the next morning, we still
>>>>>>can
>>>>>>see all data from the database, but the data from the proxy tables
>>>>>>does
>>>>>>not
>>>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that
>>>>>>the
>>>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>>>
>>>>>>When we go to Sybase Central we can see the data in the proxy tables!
>>>>>>Only
>>>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>>>
>>>>>>What can we do? Is that a bug?
>>>>>>
>>>>>>Please HELP!
>>>>>>
>>>>>>Thanks
>>>>>>Markus
>>>>>>
>>>>>
>>>>> --
>>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>>
>>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>>> breck.carter@risingroad.com
>>>>
>>>
>>> --
>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>
>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>> breck.carter@risingroad.com
>>
>>
>
>


"Stephen Rice [Sybase]" Posted on 2008-10-27 21:33:54.0Z
From: "Stephen Rice [Sybase]" <srice[at]sybase[dot]com>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub> <4905d509$1@forums-1-dub> <49061e65@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 244
Organization: Sybase
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <490633c2@forums-1-dub>
Date: 27 Oct 2008 13:33:54 -0800
X-Trace: forums-1-dub 1225143234 10.22.241.152 (27 Oct 2008 13:33:54 -0800)
X-Original-Trace: 27 Oct 2008 13:33:54 -0800, vip152.sybase.com
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:208
Article PK: 866652

I may be sorry but I'll wade in...

For answering the "timeout" question, the next step is to get communication
logs to determine why the connection is gone and it sounds like that is
under way. The proxy request actually does go "out onto the network" and
looks like a tcp/ip request to the server.

The problem you are experiencing is a much more general one for which there
is currently no magic solution that I am aware of.

The simplest sequence is (starting from no connection):
1) select 1 from proxy table which succeeds (internally we make connection,
execute sql, mark connection as good)
2) connection "goes away" for any number of reasons (internally this is not
detected)
3) select 1 from proxy table gets an error (internally we execute sql, get
communication error, connection is now marked bad)
4) select 1 from proxy table which succeeds (internally we make connection,
execute sql, mark connection as good)

AFAIK your code needs to be aware this situation can occur and handle it. I
have an internal system that uses proxy tables and I had to add the logic to
deal with the network outages, timeouts, random cosmic particles and bad
karma that infest my world. In my case I use the "select 1 from proxy
table" before I start processing from the proxy. Mine is a batch process so
this is relatively cheap.

There is an enhancement request open asking that we try and address this (I
opened it) but my understanding is the problem is a difficult one to solve
in the code.
--
/steve
Stephen Rice, Sybase iAnywhere
Please reply only to the newsgroup

iAnywhere Developer Community : http://www.ianywhere.com/developer
iAnywhere Docs : http://www.ianywhere.com/developer/product_manuals
ASA Patches and EBFs : http://downloads.sybase.com/swd/base.do
-> Choose SQL Anywhere Studio
-> Set filter to "Display ALL platforms IN ALL MONTHS"

"Markus KARG" <karg@quipsy.de> wrote in message
news:49061e65@forums-1-dub...
> Jason,
>
> thanks for your kind help. :-)
>
>> If setting the connection timeout to 0 (never) is working for you, have
>> you considered/tried setting it to a more realistic value?
>
> Well, what is a realistic value? Since the client is an application server
> pooling connections, we just don't know how long the connection is used or
> unused. The application server in fact is configured to close any unused
> connections after five minutes, so we are upset that the BASE tables can
> be read without any problems, while the PROXY tables cannot.
>
>> If the connections are timing out overnight, perhaps increasing the
>> timeout value to 12 hours would be enough to resolve the issue in most
>> cases and avoid the other problems of having no timeout set.
>
> Guide did not tell us to do so, so we abstained from changing his
> decision. But your tip sounds good. Maybe we will try to do so. My fear is
> that changing the value from 240 minutes to 12 hours might just shift the
> problem into the future instead of completely solving it.
>
> What we are interested in is a clear explanation what the cause is, and a
> 100% correctly working solution. Still nobody could tell us why BASE
> tables can be read, but PROXY tables produce connection timeout. As I
> said, it is a connection from and to the same instance of dbsrv11. So
> either both time out, or none -- or the proxy should be reconnect
> automatically.
>
> Regards
> Markus
>
>>
>> --
>> Jason Hinsperger
>> Product Manager
>> Sybase iAnywhere
>> http://iablog.sybase.com/hinsperg
>>
>> SQL Anywhere Developer Community
>> http://www.ianywhere.com/developer
>>
>>
>> "Markus KARG" <karg@quipsy.de> wrote in message
>> news:49058e31$1@forums-1-dub...
>>> Dear SQL Anywhere Professionals,
>>>
>>> I am in contact with Guido Farrenberg from Sybase Germany and we do not
>>> see a solution for the problem, still. The problem is:
>>>
>>> * Application opens JDBC connection to dbsrv11 server (in fact,
>>> "application" is GlassFish Application Server provided by Sun and
>>> Oracle).
>>> * dbsrv11 server in turn opens connection to itself for accessing the
>>> proxy table hosted in another database on the same dbsrv11 instance.
>>> * Both BASE and PROXY tables get used a lot without problems.
>>> * Then, over night, when nobody works with the application, no use of
>>> BASE or PROXY tables occurs.
>>> * Next morning, at first use, application does SELECT * FROM
>>> MyProxyTable to detect whether the pooled connection sill works and gets
>>> NO error message from the JDBC driver (Guido assumes that the
>>> PrepareStatement command is executed correctly even on a broken
>>> connection).
>>> * Then, reading data from BASE tables works well -- the connection to
>>> the "front end" server is working well.
>>> * But reading data from PROXY tables crash -- the connection from the
>>> "front end" server to the "back end" answering the proxy requests is
>>> down.
>>>
>>> WHY???
>>>
>>> Guido says there is no solution.
>>>
>>> I must set idle timeout to zero for the "internal" connection. Certainly
>>> that prevents the loss of the "internal" connection from "front end"
>>> server to "back end" server, but it leads to a lot of side effects --
>>> starting from lots of "zombie" connections kept til doomsday, to
>>> potentially never removed locks in case of crashed connections etc.
>>>
>>> So what we are seeking for desparately (!!!) is a *real* solution. From
>>> the user's view, this is clearly a bug in dbsrv11: As long as BASE
>>> tables can be read, also PROXY tables must be open for reading over THE
>>> SAME CONNECTION -- the "internal" connection must not timeout before
>>> the "original" connection is running into timeout, obviously. No
>>> application will ever understand why a SELECT on the BASE table is
>>> working, while a PROXY table "suddenly" returns "Connection lost" at the
>>> same time, OVER THE SAME CONNECTION.
>>>
>>> Please, can you provide a solution for that? Our customers are near to
>>> kill me. :-( What to do???
>>>
>>> Thanks
>>> Markus
>>>
>>> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>> Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...
>>>> Please show us the exact error message.
>>>>
>>>> Is it the ODBC proxy connection from one database to the other that is
>>>> being disconnected, or the "primary" JDBC connection from the outside
>>>> world into the first database?
>>>>
>>>> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>
>>>>>Breck,
>>>>>
>>>>>it is THE SAME server (same instance of dbsrv11.exe on the same host),
>>>>>but
>>>>>just two databases! The ODBC DSN and the dbsrv11.exe uses only
>>>>>defaults, we
>>>>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>>>>DSN)!
>>>>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>>>>should
>>>>>be transparent to the client's connection, obviously). Also, the
>>>>>default is
>>>>>240 minutes, but we encountered the problem after just 60 minutes.
>>>>>Maybe a
>>>>>bug? BTW, we had similar (but less often) problems back in SQL Anywhere
>>>>>8
>>>>>and JOnAS Application Server, so it is not a problem of the Application
>>>>>Server as it seems. And: This problems occurs at a lot of our customers
>>>>>LANs, not only in our lab. For me, it looks like a bug. I mean, when
>>>>>using
>>>>>defaults then a proxy connection on the same host should never die,
>>>>>obviously, or at least it should automatically reconnect... And no, the
>>>>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>>>>utilizing more than 50% of their possible load.
>>>>>
>>>>>Anything you want me to check?
>>>>>
>>>>>Thanks!
>>>>>Markus
>>>>>
>>>>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>>>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>>>>> It might be an idle timeout issue on the proxy connection between the
>>>>>> two servers which is probably using ODBC; see this connection
>>>>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>>>>> command:
>>>>>>
>>>>>> SQL Anywhere Server - Database Administration » Starting and
>>>>>> Connecting to Your Database » Connection parameters and network
>>>>>> protocol options » Connection parameters » Idle connection parameter
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>>>
>>>>>>>We are using GlassFishv2 Application Server to connect to a SQL Any
>>>>>>>11
>>>>>>>database (using iAnywhere JDBC driver). This connection works well.
>>>>>>>We see
>>>>>>>all data from data database. That database has proxy tables hosted by
>>>>>>>a
>>>>>>>second SQL Any 11 database on the same database server engine.
>>>>>>>GlassFish
>>>>>>>can
>>>>>>>see that data, too (obivously).
>>>>>>>
>>>>>>>We do not query any data over the night. At the next morning, we
>>>>>>>still can
>>>>>>>see all data from the database, but the data from the proxy tables
>>>>>>>does
>>>>>>>not
>>>>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is that
>>>>>>>the
>>>>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>>>>
>>>>>>>When we go to Sybase Central we can see the data in the proxy tables!
>>>>>>>Only
>>>>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>>>>
>>>>>>>What can we do? Is that a bug?
>>>>>>>
>>>>>>>Please HELP!
>>>>>>>
>>>>>>>Thanks
>>>>>>>Markus
>>>>>>>
>>>>>>
>>>>>> --
>>>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>>>
>>>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>>>> breck.carter@risingroad.com
>>>>>
>>>>
>>>> --
>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>
>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>> breck.carter@risingroad.com
>>>
>>>
>>
>>
>
>


Markus KARG Posted on 2008-10-28 06:42:03.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub> <4905d509$1@forums-1-dub> <49061e65@forums-1-dub> <490633c2@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 272
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4906b43b@forums-1-dub>
Date: 27 Oct 2008 22:42:03 -0800
X-Trace: forums-1-dub 1225176123 10.22.241.152 (27 Oct 2008 22:42:03 -0800)
X-Original-Trace: 27 Oct 2008 22:42:03 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:209
Article PK: 866653

Stephen,

you might be right that this is a general problem in SQL Anywhere, but see
that the application cannot deal with that: The application is a commercial
application server product that certainly will not be changed by its vendor
just to workaround Sybase's inability to keep open their own, internal
connections. For virtually anybody besides Sybase employees this looks 100%
like a fault in the construction of the proxy system. The solution is as
simple as: A Proxy connection must not terminate due to idle timeout as long
as the corresponding "base" connection is not terminated. Not more. Not
less. It is just a bug, but not a programming bug -- it is a construction
bug. You propose that an application shall be prepared to deal with that?
Well, how shall the application server's connection pooling sub system deal
with that?

Regards
Markus

"Stephen Rice [Sybase]" <srice[at]sybase[dot]com> schrieb im Newsbeitrag
news:490633c2@forums-1-dub...

>I may be sorry but I'll wade in...
>
> For answering the "timeout" question, the next step is to get
> communication logs to determine why the connection is gone and it sounds
> like that is under way. The proxy request actually does go "out onto the
> network" and looks like a tcp/ip request to the server.
>
> The problem you are experiencing is a much more general one for which
> there is currently no magic solution that I am aware of.
>
> The simplest sequence is (starting from no connection):
> 1) select 1 from proxy table which succeeds (internally we make
> connection, execute sql, mark connection as good)
> 2) connection "goes away" for any number of reasons (internally this is
> not detected)
> 3) select 1 from proxy table gets an error (internally we execute sql, get
> communication error, connection is now marked bad)
> 4) select 1 from proxy table which succeeds (internally we make
> connection, execute sql, mark connection as good)
>
> AFAIK your code needs to be aware this situation can occur and handle it.
> I have an internal system that uses proxy tables and I had to add the
> logic to deal with the network outages, timeouts, random cosmic particles
> and bad karma that infest my world. In my case I use the "select 1 from
> proxy table" before I start processing from the proxy. Mine is a batch
> process so this is relatively cheap.
>
> There is an enhancement request open asking that we try and address this
> (I opened it) but my understanding is the problem is a difficult one to
> solve in the code.
> --
> /steve
> Stephen Rice, Sybase iAnywhere
> Please reply only to the newsgroup
>
> iAnywhere Developer Community : http://www.ianywhere.com/developer
> iAnywhere Docs : http://www.ianywhere.com/developer/product_manuals
> ASA Patches and EBFs : http://downloads.sybase.com/swd/base.do
> -> Choose SQL Anywhere Studio
> -> Set filter to "Display ALL platforms IN ALL MONTHS"
>
> "Markus KARG" <karg@quipsy.de> wrote in message
> news:49061e65@forums-1-dub...
>> Jason,
>>
>> thanks for your kind help. :-)
>>
>>> If setting the connection timeout to 0 (never) is working for you, have
>>> you considered/tried setting it to a more realistic value?
>>
>> Well, what is a realistic value? Since the client is an application
>> server pooling connections, we just don't know how long the connection is
>> used or unused. The application server in fact is configured to close any
>> unused connections after five minutes, so we are upset that the BASE
>> tables can be read without any problems, while the PROXY tables cannot.
>>
>>> If the connections are timing out overnight, perhaps increasing the
>>> timeout value to 12 hours would be enough to resolve the issue in most
>>> cases and avoid the other problems of having no timeout set.
>>
>> Guide did not tell us to do so, so we abstained from changing his
>> decision. But your tip sounds good. Maybe we will try to do so. My fear
>> is that changing the value from 240 minutes to 12 hours might just shift
>> the problem into the future instead of completely solving it.
>>
>> What we are interested in is a clear explanation what the cause is, and a
>> 100% correctly working solution. Still nobody could tell us why BASE
>> tables can be read, but PROXY tables produce connection timeout. As I
>> said, it is a connection from and to the same instance of dbsrv11. So
>> either both time out, or none -- or the proxy should be reconnect
>> automatically.
>>
>> Regards
>> Markus
>>
>>>
>>> --
>>> Jason Hinsperger
>>> Product Manager
>>> Sybase iAnywhere
>>> http://iablog.sybase.com/hinsperg
>>>
>>> SQL Anywhere Developer Community
>>> http://www.ianywhere.com/developer
>>>
>>>
>>> "Markus KARG" <karg@quipsy.de> wrote in message
>>> news:49058e31$1@forums-1-dub...
>>>> Dear SQL Anywhere Professionals,
>>>>
>>>> I am in contact with Guido Farrenberg from Sybase Germany and we do not
>>>> see a solution for the problem, still. The problem is:
>>>>
>>>> * Application opens JDBC connection to dbsrv11 server (in fact,
>>>> "application" is GlassFish Application Server provided by Sun and
>>>> Oracle).
>>>> * dbsrv11 server in turn opens connection to itself for accessing the
>>>> proxy table hosted in another database on the same dbsrv11 instance.
>>>> * Both BASE and PROXY tables get used a lot without problems.
>>>> * Then, over night, when nobody works with the application, no use of
>>>> BASE or PROXY tables occurs.
>>>> * Next morning, at first use, application does SELECT * FROM
>>>> MyProxyTable to detect whether the pooled connection sill works and
>>>> gets NO error message from the JDBC driver (Guido assumes that the
>>>> PrepareStatement command is executed correctly even on a broken
>>>> connection).
>>>> * Then, reading data from BASE tables works well -- the connection to
>>>> the "front end" server is working well.
>>>> * But reading data from PROXY tables crash -- the connection from the
>>>> "front end" server to the "back end" answering the proxy requests is
>>>> down.
>>>>
>>>> WHY???
>>>>
>>>> Guido says there is no solution.
>>>>
>>>> I must set idle timeout to zero for the "internal" connection.
>>>> Certainly that prevents the loss of the "internal" connection from
>>>> "front end" server to "back end" server, but it leads to a lot of side
>>>> effects -- starting from lots of "zombie" connections kept til
>>>> doomsday, to potentially never removed locks in case of crashed
>>>> connections etc.
>>>>
>>>> So what we are seeking for desparately (!!!) is a *real* solution. From
>>>> the user's view, this is clearly a bug in dbsrv11: As long as BASE
>>>> tables can be read, also PROXY tables must be open for reading over THE
>>>> SAME CONNECTION -- the "internal" connection must not timeout before
>>>> the "original" connection is running into timeout, obviously. No
>>>> application will ever understand why a SELECT on the BASE table is
>>>> working, while a PROXY table "suddenly" returns "Connection lost" at
>>>> the same time, OVER THE SAME CONNECTION.
>>>>
>>>> Please, can you provide a solution for that? Our customers are near to
>>>> kill me. :-( What to do???
>>>>
>>>> Thanks
>>>> Markus
>>>>
>>>> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>>>> Newsbeitrag news:kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com...
>>>>> Please show us the exact error message.
>>>>>
>>>>> Is it the ODBC proxy connection from one database to the other that is
>>>>> being disconnected, or the "primary" JDBC connection from the outside
>>>>> world into the first database?
>>>>>
>>>>> On 7 Oct 2008 23:24:01 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>>
>>>>>>Breck,
>>>>>>
>>>>>>it is THE SAME server (same instance of dbsrv11.exe on the same host),
>>>>>>but
>>>>>>just two databases! The ODBC DSN and the dbsrv11.exe uses only
>>>>>>defaults, we
>>>>>>did not change them (no options set, neither in dbsrv11.exe nor in the
>>>>>>DSN)!
>>>>>>Also, why is it working in Sybase Central but not in JDBC (the proxy
>>>>>>should
>>>>>>be transparent to the client's connection, obviously). Also, the
>>>>>>default is
>>>>>>240 minutes, but we encountered the problem after just 60 minutes.
>>>>>>Maybe a
>>>>>>bug? BTW, we had similar (but less often) problems back in SQL
>>>>>>Anywhere 8
>>>>>>and JOnAS Application Server, so it is not a problem of the
>>>>>>Application
>>>>>>Server as it seems. And: This problems occurs at a lot of our
>>>>>>customers
>>>>>>LANs, not only in our lab. For me, it looks like a bug. I mean, when
>>>>>>using
>>>>>>defaults then a proxy connection on the same host should never die,
>>>>>>obviously, or at least it should automatically reconnect... And no,
>>>>>>the
>>>>>>number of connections is just a dozen, neither RAM nor CPU nor HDD is
>>>>>>utilizing more than 50% of their possible load.
>>>>>>
>>>>>>Anything you want me to check?
>>>>>>
>>>>>>Thanks!
>>>>>>Markus
>>>>>>
>>>>>>"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb
>>>>>>im
>>>>>>Newsbeitrag news:8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com...
>>>>>>> It might be an idle timeout issue on the proxy connection between
>>>>>>> the
>>>>>>> two servers which is probably using ODBC; see this connection
>>>>>>> parameter which you can put in the ODBC DSN or in the CREATE SERVER
>>>>>>> command:
>>>>>>>
>>>>>>> SQL Anywhere Server - Database Administration » Starting and
>>>>>>> Connecting to Your Database » Connection parameters and network
>>>>>>> protocol options » Connection parameters » Idle connection parameter
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On 7 Oct 2008 00:11:36 -0700, "Markus KARG" <karg@quipsy.de> wrote:
>>>>>>>
>>>>>>>>We are using GlassFishv2 Application Server to connect to a SQL Any
>>>>>>>>11
>>>>>>>>database (using iAnywhere JDBC driver). This connection works well.
>>>>>>>>We see
>>>>>>>>all data from data database. That database has proxy tables hosted
>>>>>>>>by a
>>>>>>>>second SQL Any 11 database on the same database server engine.
>>>>>>>>GlassFish
>>>>>>>>can
>>>>>>>>see that data, too (obivously).
>>>>>>>>
>>>>>>>>We do not query any data over the night. At the next morning, we
>>>>>>>>still can
>>>>>>>>see all data from the database, but the data from the proxy tables
>>>>>>>>does
>>>>>>>>not
>>>>>>>>occur. Instead the JDBC Driver throws a SQL Exception telling is
>>>>>>>>that the
>>>>>>>>foreign server cannot be reached: "Connection terminated" it says.
>>>>>>>>
>>>>>>>>When we go to Sybase Central we can see the data in the proxy
>>>>>>>>tables! Only
>>>>>>>>the JDBC driver is not seeting that data. VERY STRANGE!
>>>>>>>>
>>>>>>>>What can we do? Is that a bug?
>>>>>>>>
>>>>>>>>Please HELP!
>>>>>>>>
>>>>>>>>Thanks
>>>>>>>>Markus
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>>>>
>>>>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>>>>> breck.carter@risingroad.com
>>>>>>
>>>>>
>>>>> --
>>>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>>>
>>>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>>>> breck.carter@risingroad.com
>>>>
>>>>
>>>
>>>
>>
>>
>
>


Jeff Albion (Sybase iAnywhere) Posted on 2008-10-27 14:19:00.0Z
From: "Jeff Albion (Sybase iAnywhere)" <firstname.lastname@ianywhere.com>
User-Agent: Thunderbird 2.0.0.17 (Windows/20080914)
MIME-Version: 1.0
Newsgroups: sybase.public.sqlanywhere
Subject: Re: Poxy table connection terminated?!
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub>
In-Reply-To: <49058e31$1@forums-1-dub>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4905cdd4$1@forums-1-dub>
Date: 27 Oct 2008 06:19:00 -0800
X-Trace: forums-1-dub 1225117140 10.22.241.152 (27 Oct 2008 06:19:00 -0800)
X-Original-Trace: 27 Oct 2008 06:19:00 -0800, vip152.sybase.com
Lines: 37
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:204
Article PK: 866649

Markus,

Since you're already working with Guido, it makes the most sense for you
to continue working with him on this case. Guido can directly involve
engineering groups once we've captured the observed behaviour in-house.

What types of debugging steps have you already tried with Guido? Do we
know what state the internal connection is in? Have you checked to see
what the "CIS" logging is showing? (See:
http://dcx.sybase.com/1100en/dbadmin_en11/dboptions-s-5821452.html ).
Have you provided Guido a reproducible so that engineering can take a
look at the observed behaviour and help us determine if there's a
correct solution for this issue?

Just from this newsgroup thread, there are several things that are not
clear:

- How does Glassfish perform connection maintenance from its side? Are
the connections pooled? What happens if you turn off connection pooling
in Glassfish?
- What's your remote server definition?
- Exact version of v11? (dbsrv11 -v)

I'd highly recommend collecting this information and/or provide a
reproducible to Guido so we can best address your issue.

Regards,

--
Jeff Albion, SQL Anywhere Engineering
Sybase iAnywhere

iAnywhere Developer Community :
http://www.sybase.com/developer/library/sql-anywhere-techcorner
iAnywhere Documentation : http://www.ianywhere.com/developer/product_manuals
SQL Anywhere Patches and EBFs :
http://downloads.sybase.com/swd/summary.do?baseprod=144&client=ianywhere&timeframe=0


Markus KARG Posted on 2008-10-27 19:58:30.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub> <4905cdd4$1@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 101
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <49061d66$1@forums-1-dub>
Date: 27 Oct 2008 11:58:30 -0800
X-Trace: forums-1-dub 1225137510 10.22.241.152 (27 Oct 2008 11:58:30 -0800)
X-Original-Trace: 27 Oct 2008 11:58:30 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:206
Article PK: 866650

Jeff,

> Since you're already working with Guido, it makes the most sense for you
> to continue working with him on this case. Guido can directly involve
> engineering groups once we've captured the observed behaviour in-house.

Well, what Guido said sounded to me like he has no more idea what he shall
do. If I understood his explanations correctly then actually he is about to
stop any further investigations with the justification that setting idle
timeout to zero would be an acceptable solution (but I do not accept that
solution).

> What types of debugging steps have you already tried with Guido?

He asked me to check whether idle timeout = 0 for the proxy connection would
prevent connection loss. I told him that it does. But certainly that is not
the solution (but he thinks that applying idle timeout = 0 IS the final
solution -- and ignores the side effects that this will imply).

Then he asked me to provide a "-zr all" log. I am currently preparing that
log.

> Do we know what state the internal connection is in?

No. How to find out?

> Have you checked to see what the "CIS" logging is showing? (See:
> http://dcx.sybase.com/1100en/dbadmin_en11/dboptions-s-5821452.html ).

He did not tell me anything about that option, so we did not enable this. He
just told me to enable "-zr all".

> Have you provided Guido a reproducible so that engineering can take a look
> at the observed behaviour and help us determine if there's a correct
> solution for this issue?

I gave Guide the same information that I gave to you, so he should be able
to reproduce the problem, as it is really simple.

> Just from this newsgroup thread, there are several things that are not
> clear:
>
> - How does Glassfish perform connection maintenance from its side?

As I told Guide, it seems GlassFish does a JDBC prepareStatement on "SELECT
* FROM MyProxyTable" and checks for SQLException. Actually I can only
configure the table, so I provided MyProxyTable to ensure that the Proxy is
involved, but I do not know what the actual SQL command is or whether any
data is really read.

> Are the connections pooled?

Yes. The pool is adjusted to close any unused connections after five
minutes -- hours before the idle timeout happens.

> What happens if you turn off connection pooling in Glassfish?

Connection pooling cannot be turned off in GlassFish.

> - What's your remote server definition?

Server Type: SQL Anywhere
Connection Type: ODBC
Connection Info: RH/FMEA

(RH/FMEA is the ODBC DSN of the proxy server, which is the same instance of
dbsrv11 in our case)

> - Exact version of v11? (dbsrv11 -v)

11.0.0.1264

> I'd highly recommend collecting this information and/or provide a
> reproducible to Guido so we can best address your issue.

Already done, as I said. But it seems Guido actually does not want to
reproduce this on his machine... :-( I told him that lots of our customers
have the same problems, and had that back in ASA 8.0.3 times already. So it
should be reproducible for anybody everywhere anythime (at least we could at
once after customers told us, even in a VMware virtual machine).

Thanks a lot for your kind help! :-)

Regards
Markus

>
> Regards,
>
> --
> Jeff Albion, SQL Anywhere Engineering
> Sybase iAnywhere
>
> iAnywhere Developer Community :
> http://www.sybase.com/developer/library/sql-anywhere-techcorner
> iAnywhere Documentation :
> http://www.ianywhere.com/developer/product_manuals
> SQL Anywhere Patches and EBFs :
> http://downloads.sybase.com/swd/summary.do?baseprod=144&client=ianywhere&timeframe=0


Markus KARG Posted on 2008-10-28 07:00:44.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <8clme4lie11rmha6j7na5i2o6lbjfcpobn@4ax.com> <48ec5201$1@forums-1-dub> <kg8pe41lrujnaph40jdkqr7ovqk520967o@4ax.com> <49058e31$1@forums-1-dub> <4905cdd4$1@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 94
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4906b89c@forums-1-dub>
Date: 27 Oct 2008 23:00:44 -0800
X-Trace: forums-1-dub 1225177244 10.22.241.152 (27 Oct 2008 23:00:44 -0800)
X-Original-Trace: 27 Oct 2008 23:00:44 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:210
Article PK: 866656

Jeff,

thanks for your kind help. Today I have more information to you, as I have
created "-zr all" logs.

The problem is really as simple as I told in the newsgroup already: My
application regularly SELECTs from a BASE table even at night (I forgot to
mention), so the BASE connections all keep opened and "non-idle" for longer
than a few minutes. Also, before using the BASE connection, the application
server does a PREPARE on SELECT * FROM MyProxyTable to ensure everything
still is running. And the server gets an "OK, all is fine" back from the
driver -- while actually the connection is already terminated! And in fact,
the proxy connection gets terminated due to the sole fact that there is no
traffic on the PROXY tables (while there IS traffic on the BASE tables). In
the morning, when we want to access the PROXY table, another PREPARE (OK)
and FETCH (connection terminated) is done and you see the mess...

Neither can I make Oracle to change their application server, nor can I make
my 100% Java EE compliant application to deal with that (what shall it do?
It uses JPA to find objects, and Oracle throws an SQLException telling me
that the connection is open but the table cannot get read -- virtually every
application will fail in the assumption that the database server is broken).

So to me (and virtually everybody besides Sybase employees) this behaviour
looks like a "construction fault". It just makes no sense to terminate a
PROXY connection while the corresponding base connection still is not
terminated. It EVER could happen that a connection that did not use a PROXY
table for hours will suddenly access that PROXY table. It is just FALSE to
deal with proxy connections the same way that "normal" connections are dealt
with. Or to say it with other words: If you allow the proxy connection to
disconnect while the base connection is connected, then you also must
reconnect when the base connection accesses the proxy table later! Either -
Or!

From my understanding there is so far no solution besides changing the
application server or changing my Java EE application. Well the first is
something that Oracle won't do. The second is something that I must not
do -- since we are not allowed to add database vendor specific workarounds
to our code, as we are and ISV.

So do you have any trick left over? I mean, the current situation is: Our
over 1200 enterprise customers cannot use SQL Anywhere, since it does not
correctly support connection pooling! :-(

Any ideas -- besides doing programmatic changes to my application or the
application server -- are highly appreciated!!! :-)

Thanks a lot! :-)
Markus



"Jeff Albion (Sybase iAnywhere)" <firstname.lastname@ianywhere.com> schrieb
im Newsbeitrag news:4905cdd4$1@forums-1-dub...

> Markus,
>
> Since you're already working with Guido, it makes the most sense for you
> to continue working with him on this case. Guido can directly involve
> engineering groups once we've captured the observed behaviour in-house.
>
> What types of debugging steps have you already tried with Guido? Do we
> know what state the internal connection is in? Have you checked to see
> what the "CIS" logging is showing? (See:
> http://dcx.sybase.com/1100en/dbadmin_en11/dboptions-s-5821452.html ). Have
> you provided Guido a reproducible so that engineering can take a look at
> the observed behaviour and help us determine if there's a correct solution
> for this issue?
>
> Just from this newsgroup thread, there are several things that are not
> clear:
>
> - How does Glassfish perform connection maintenance from its side? Are the
> connections pooled? What happens if you turn off connection pooling in
> Glassfish?
> - What's your remote server definition?
> - Exact version of v11? (dbsrv11 -v)
>
> I'd highly recommend collecting this information and/or provide a
> reproducible to Guido so we can best address your issue.
>
> Regards,
>
> --
> Jeff Albion, SQL Anywhere Engineering
> Sybase iAnywhere
>
> iAnywhere Developer Community :
> http://www.sybase.com/developer/library/sql-anywhere-techcorner
> iAnywhere Documentation :
> http://www.ianywhere.com/developer/product_manuals
> SQL Anywhere Patches and EBFs :
> http://downloads.sybase.com/swd/summary.do?baseprod=144&client=ianywhere&timeframe=0


Karim Khamis Posted on 2008-10-28 16:54:40.0Z
From: Karim Khamis <kkhamis@sybase.com>
User-Agent: Thunderbird 2.0.0.17 (Windows/20080914)
MIME-Version: 1.0
Newsgroups: sybase.public.sqlanywhere
Subject: Re: Poxy table connection terminated?!
References: <48eb0ba8$1@forums-1-dub>
In-Reply-To: <48eb0ba8$1@forums-1-dub>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <490743d0$1@forums-1-dub>
Date: 28 Oct 2008 08:54:40 -0800
X-Trace: forums-1-dub 1225212880 10.22.241.152 (28 Oct 2008 08:54:40 -0800)
X-Original-Trace: 28 Oct 2008 08:54:40 -0800, vip152.sybase.com
Lines: 76
X-Authenticated-User: techsupp
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:211
Article PK: 866655

Markus,

We have in the past considered the option of having remote connections
automatically reconnect after a timeout or any other reason; BUT, doing
so can lead to some very serious problems. Consider the following scenario:

An application starts up and updates a remote table. The application
then runs for several hours updating local tables. During this time the
remote connection times out. When the application is all done updating
the local tables, it executes a query that involves a remote table on
the same remote server as the first remote update took place. Now
suppose at this time the server automatically reconnected to the remote
server and sent a warning back to the application indicating that it
reconnected. Warnings often go ignored or unnoticed, so the application
probably will decide all is well and issue a commit at this point. The
effect of all this is that the local updates will all get committed
while the remote update is completely lost (it got rolled back when the
idle timeout occurred). If the above scenario takes place every night
when no human is around, then the problem might go unnoticed for days,
weeks or even months. You can just imagine the problem the customer will
have if and when he/she does notice.

The other point is that the back end may not be another SA
server/database. The remote may be Oracle, DB2, ASE, SQL Server, Lotus,
Excel, Access, or any other application that has an ODBC driver. As a
result, the SA server cannot rely on any mechanism to ensure that the
remote connection will always stay open. In fact, in many cases, the
server cannot even poll regularly to ensure the connection is still open
at any given time.

The only solution is the one that Guido and others have already
suggested, and that is to set the idle timeout. Note that you need not
set idle timeout on a global scale. If you are using a DSN to connect to
the remote server, then setting the idle timeout within the DSN will
only affect those connections that use that DSN. If, on the other hand,
you are not using a DSN but are instead providing a full connection
string in the USING clause of the remote server, then you can set the
idle timeout within that USING clause and only those remote connections
that the sever creates to that specific remote server will be affected.
Also, using idle timeout in this situation will not leave any orphaned
connections because the remote connections will automatically get
dropped as soon as the local SA connection is dropped.

You might also want to look at the documentation for the ALTER SERVER
command. You can use the ALTER SERVER command to drop remote connections
to any remote server. Once you explicitly drop a remote connection, the
very next access to the remote server will result in a reconnect. The
difference here is that you, the user, has full control if/when the
reconnect takes place and with the full understanding of what might have
gone wrong in the meantime.

Hope that helps,
Karim

Markus KARG wrote:
> We are using GlassFishv2 Application Server to connect to a SQL Any 11
> database (using iAnywhere JDBC driver). This connection works well. We see
> all data from data database. That database has proxy tables hosted by a
> second SQL Any 11 database on the same database server engine. GlassFish can
> see that data, too (obivously).
>
> We do not query any data over the night. At the next morning, we still can
> see all data from the database, but the data from the proxy tables does not
> occur. Instead the JDBC Driver throws a SQL Exception telling is that the
> foreign server cannot be reached: "Connection terminated" it says.
>
> When we go to Sybase Central we can see the data in the proxy tables! Only
> the JDBC driver is not seeting that data. VERY STRANGE!
>
> What can we do? Is that a bug?
>
> Please HELP!
>
> Thanks
> Markus
>
>


Markus KARG Posted on 2008-10-29 07:34:35.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <490743d0$1@forums-1-dub>
Subject: Re: Poxy table connection terminated?!
Lines: 83
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4908120b@forums-1-dub>
Date: 28 Oct 2008 23:34:35 -0800
X-Trace: forums-1-dub 1225265675 10.22.241.152 (28 Oct 2008 23:34:35 -0800)
X-Original-Trace: 28 Oct 2008 23:34:35 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:212
Article PK: 866654


> probably will decide all is well and issue a commit at this point. The
> effect of all this is that the local updates will all get committed while
> the remote update is completely lost (it got rolled back when the idle
> timeout occurred).

While I also like more the idea of not truncate the connection instead of
reconnect it, the solution to your above mentioned problem is quite simple:
Do not glue the lifetime of the transaction to the lifetime of the
connection. In other words, do not rollback at disconnect. I know this is
not how SQL Anywhere works, but it solves that problem.

> The other point is that the back end may not be another SA
> server/database. The remote may be Oracle, DB2, ASE, SQL Server, Lotus,
> Excel, Access, or any other application that has an ODBC driver. As a
> result, the SA server cannot rely on any mechanism to ensure that the
> remote connection will always stay open. In fact, in many cases, the
> server cannot even poll regularly to ensure the connection is still open
> at any given time.

I only was talking about connections to SA, and the SA server knows very
well whether the type of the target: SA.

> The only solution is the one that Guido and others have already suggested,
> and that is to set the idle timeout.

I have to disagree. First, a much simple solution was found by Guide
yesterday: Don't use TCP/IP but use SharedMemory. There is no idle timeout,
so the proxy connections just do not disconnect when not use. So it is
untrue that there is only ONE solution. Second, setting idle timeout to zero
can potentially result is keeping locks forever, as Guide said. So that is
not a solution, but just a workaround. The real solution is that SA learns
the difference between PROXY and BASE connections: If SA opens a proxy
connection to a remote SA server, it must not terminate THIS connection at
client idleness but solely at close command or at lifeness timeout -- which
is a difference to setting IDLE TIMEOUT to zero at the DSN, since it only
applies to this single connection instead to all connections done by any
other application using the same DSN. Certainly the user could workaround by
supplying different DSN using different IDLE TIMEOUT values. But it stays
what it is, a workaround. If SA knows that it connects to SA, why doesn't it
just send the IDLE TIMEOUT = 0 on it's own? It knows it cannot handle the
broken connection, so why not dealing with that on it's own? And why not
closing the original connection in case it recognizes a timeout on it's
proxy, so BASE and PROXY behaves the same? THAT are solutions. All the rest
are workarounds.


> Also, using idle timeout in this situation will not leave any orphaned
> connections because the remote connections will automatically get dropped
> as soon as the local SA connection is dropped.

That is the problem: The local SA connection IS NOT GETTING DROPPED since
all the BASE TABLE selects are still working. All I want is that the base
tables and the proxy tables behaves the same. If the proxy is down, the base
shall be down, too. No application will understand that it is NOT a server
outage that tables behave differently inside of the same connection. Got the
point? No need to answer BASE selects correctly at that moment.

> You might also want to look at the documentation for the ALTER SERVER
> command. You can use the ALTER SERVER command to drop remote connections
> to any remote server. Once you explicitly drop a remote connection, the
> very next access to the remote server will result in a reconnect.

Okay, tell me how to tell an application server how to use that feature? We
are an ISV, writing non-Sybase-aware Java EE applications, running on a
non-Sybase-aware application server. So how to apply that command?

> Hope that helps,

Unfortunately not. So we just switched to SharedMemory now, but the problem
will be back as soon as the remote server is really remote. As long as the
original connection is working "partially successful" in case a proxy
connection timemout occured, virtually any non-Sybase-aware application will
fail. And the solution could be so simple. If Sybase connects to Sybase,
send the IDLE TIMEOUT internally but do not bother the DSN with it so other
applications are not affected. If Sybase connects NOT to Sybase, drop the
connection as soon as the proxy connection drops. It's just as simple, and
would solve all problems. So why not doing it?

Regards
Markus


Breck Carter [TeamSybase] Posted on 2008-10-29 09:54:13.0Z
From: "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com>
Newsgroups: sybase.public.sqlanywhere
Subject: Re: Poxy table connection terminated?!
Organization: RisingRoad Professional Services
Reply-To: NOSPAM__breck.carter@gmail.com
Message-ID: <ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com>
References: <48eb0ba8$1@forums-1-dub> <490743d0$1@forums-1-dub> <4908120b@forums-1-dub>
X-Newsreader: Forte Agent 2.0/32.640
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Date: 29 Oct 2008 01:54:13 -0800
X-Trace: forums-1-dub 1225274053 10.22.241.152 (29 Oct 2008 01:54:13 -0800)
X-Original-Trace: 29 Oct 2008 01:54:13 -0800, vip152.sybase.com
Lines: 16
X-Authenticated-User: TeamSybase
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:213
Article PK: 866657


On 28 Oct 2008 23:34:35 -0800, "Markus KARG" <karg@quipsy.de> wrote:

>do not rollback at disconnect

Sorry, that's not acceptable. The fact a connection is being dropped
is not proof that a (possibly incomplete) transaction should be
committed. It's not just "how SQL Anywhere works", it is what makes
sense.

Breck

--
Breck Carter http://sqlanywhere.blogspot.com/

RisingRoad SQL Anywhere and MobiLink Professional Services
breck.carter@risingroad.com


Markus KARG Posted on 2008-10-29 13:42:57.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <490743d0$1@forums-1-dub> <4908120b@forums-1-dub> <ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com>
Subject: Re: Poxy table connection terminated?!
Lines: 31
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <49086861$1@forums-1-dub>
Date: 29 Oct 2008 05:42:57 -0800
X-Trace: forums-1-dub 1225287777 10.22.241.152 (29 Oct 2008 05:42:57 -0800)
X-Original-Trace: 29 Oct 2008 05:42:57 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:214
Article PK: 866659

I did not say that the transaction shall be committed. I said it shall not
rollback. In other words: do not terminate the connection at all, but split
transaction termination from connection termination. That way, uncommitted
transactions will survive reconnects. The JDBC specification does not say
that a transaction must be terminated when the connection is lost. I do not
see any problem here besides the fact that it is not the way SQL Anyhwhere
works like.

Regards
Markus

"Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
Newsbeitrag news:ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com...

> On 28 Oct 2008 23:34:35 -0800, "Markus KARG" <karg@quipsy.de> wrote:
>
>>do not rollback at disconnect
>
> Sorry, that's not acceptable. The fact a connection is being dropped
> is not proof that a (possibly incomplete) transaction should be
> committed. It's not just "how SQL Anywhere works", it is what makes
> sense.
>
> Breck
>
> --
> Breck Carter http://sqlanywhere.blogspot.com/
>
> RisingRoad SQL Anywhere and MobiLink Professional Services
> breck.carter@risingroad.com


"Frank Ploessel" <fpl... Posted on 2008-10-29 15:46:38.0Z
Subject: Re: Poxy table connection terminated?!
From: "Frank Ploessel" <fpl...@d_e.i_m_s_h_e_a_l_t_h.c_o_m>
Content-Type: text/plain; format=flowed; delsp=yes; charset=iso-8859-15
MIME-Version: 1.0
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <490743d0$1@forums-1-dub> <4908120b@forums-1-dub> <ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com> <49086861$1@forums-1-dub>
Content-Transfer-Encoding: 7bit
Message-ID: <op.ujsl30kpj0bybf@bonw01164.internal.imsglobal.com>
User-Agent: Opera Mail/9.51 (Win32)
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Date: 29 Oct 2008 07:46:38 -0800
X-Trace: forums-1-dub 1225295198 10.22.241.152 (29 Oct 2008 07:46:38 -0800)
X-Original-Trace: 29 Oct 2008 07:46:38 -0800, vip152.sybase.com
Lines: 46
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:215
Article PK: 866661

Markus,

Do you know any example of another DBMS that supports transactions that
survive disconnect/reconnect?

Frank

Am 29.10.2008, 14:42 Uhr, schrieb Markus KARG <karg@quipsy.de>:

> I did not say that the transaction shall be committed. I said it shall
> not
> rollback. In other words: do not terminate the connection at all, but
> split
> transaction termination from connection termination. That way,
> uncommitted
> transactions will survive reconnects. The JDBC specification does not say
> that a transaction must be terminated when the connection is lost. I do
> not
> see any problem here besides the fact that it is not the way SQL
> Anyhwhere
> works like.
>
> Regards
> Markus
>
> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
> Newsbeitrag news:ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com...
>> On 28 Oct 2008 23:34:35 -0800, "Markus KARG" <karg@quipsy.de> wrote:
>>
>>> do not rollback at disconnect
>>
>> Sorry, that's not acceptable. The fact a connection is being dropped
>> is not proof that a (possibly incomplete) transaction should be
>> committed. It's not just "how SQL Anywhere works", it is what makes
>> sense.
>>
>> Breck
>>
>> --
>> Breck Carter http://sqlanywhere.blogspot.com/
>>
>> RisingRoad SQL Anywhere and MobiLink Professional Services
>> breck.carter@risingroad.com
>
>


Markus KARG Posted on 2008-10-30 08:26:18.0Z
From: "Markus KARG" <karg@quipsy.de>
Newsgroups: sybase.public.sqlanywhere
References: <48eb0ba8$1@forums-1-dub> <490743d0$1@forums-1-dub> <4908120b@forums-1-dub> <ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com> <49086861$1@forums-1-dub> <op.ujsl30kpj0bybf@bonw01164.internal.imsglobal.com>
Subject: Re: Poxy table connection terminated?!
Lines: 69
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <49096faa@forums-1-dub>
Date: 30 Oct 2008 00:26:18 -0800
X-Trace: forums-1-dub 1225355178 10.22.241.152 (30 Oct 2008 00:26:18 -0800)
X-Original-Trace: 30 Oct 2008 00:26:18 -0800, vip152.sybase.com
X-Authenticated-User: panorama
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.sqlanywhere:216
Article PK: 866658

I once worked with one, but forgot it's name and vendor unfortunately. There
it was the case that I had to manually begin and commit transactions. After
connecting, no transaction existed. So there was a separation of connect /
disconnect and begin trans / commit trans. I remembered about this because
it was so surprising that after the connect I had to manually start the TX.

Also, isn't the XA enforcing that? I mean, in XA the start and end of a
transaction is coordinated by a central server. So despite the existing
connection, there might be no existing transaction. And the transaction can
live while there is no current connection. I know, SA doesn't support XA.
But it is a sample of separation of connection and transaction.

Anyways, the idea of the reconnect was just an example of a possible
solution. My favorite would be no to cut the proxy connection as long as the
base connection lives. This would be far more transparent to the caller.

Regards
Markus

"Frank Ploessel" <fpl...@d_e.i_m_s_h_e_a_l_t_h.c_o_m> schrieb im Newsbeitrag
news:op.ujsl30kpj0bybf@bonw01164.internal.imsglobal.com...

> Markus,
>
> Do you know any example of another DBMS that supports transactions that
> survive disconnect/reconnect?
>
> Frank
>
> Am 29.10.2008, 14:42 Uhr, schrieb Markus KARG <karg@quipsy.de>:
>
>> I did not say that the transaction shall be committed. I said it shall
>> not
>> rollback. In other words: do not terminate the connection at all, but
>> split
>> transaction termination from connection termination. That way,
>> uncommitted
>> transactions will survive reconnects. The JDBC specification does not say
>> that a transaction must be terminated when the connection is lost. I do
>> not
>> see any problem here besides the fact that it is not the way SQL
>> Anyhwhere
>> works like.
>>
>> Regards
>> Markus
>>
>> "Breck Carter [TeamSybase]" <NOSPAM__breck.carter@gmail.com> schrieb im
>> Newsbeitrag news:ejcgg4ts0rq8mva0bkuv5g4n0inmfvdif1@4ax.com...
>>> On 28 Oct 2008 23:34:35 -0800, "Markus KARG" <karg@quipsy.de> wrote:
>>>
>>>> do not rollback at disconnect
>>>
>>> Sorry, that's not acceptable. The fact a connection is being dropped
>>> is not proof that a (possibly incomplete) transaction should be
>>> committed. It's not just "how SQL Anywhere works", it is what makes
>>> sense.
>>>
>>> Breck
>>>
>>> --
>>> Breck Carter http://sqlanywhere.blogspot.com/
>>>
>>> RisingRoad SQL Anywhere and MobiLink Professional Services
>>> breck.carter@risingroad.com
>>
>>
>