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.

Problems after Metadata descripor error

12 posts in General Discussion Last posting was on 2010-07-08 16:38:48.0Z
jbuhl Posted on 2010-06-29 19:01:33.0Z
Sender: 3641.4c2a3906.1804289383@sybase.com
From: jbuhl
Newsgroups: sybase.public.ase.general
Subject: Problems after Metadata descripor error
X-Mailer: WebNews to Mail Gateway v1.1t
Message-ID: <4c2a430d.384e.1681692777@sybase.com>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 29 Jun 2010 12:01:33 -0700
X-Trace: forums-1-dub 1277838093 10.22.241.41 (29 Jun 2010 12:01:33 -0700)
X-Original-Trace: 29 Jun 2010 12:01:33 -0700, 10.22.241.41
Lines: 39
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29327
Article PK: 78561

ASE 15.0.3 ESD #2
Solaris 10

Server started refusing application logins. DBA logs in via
isql and gets connected fine but when he issues "use
database.." server disconnects his session without message.
DBA logs back in and immediately issues shutdown command and
gets message waiting for process to complete. He then logs
in via dbartisan gui admin tool. This tool pops serveral
error screens (did not record) and then settles in and he is
able to see a list of processes; one of which is running a
select operation. He kills this process via the GUI admin
tool and the server shuts down clean. He restarts ASE
server and everything has been fine for a couple days.
This happened on a Sunday afternoon which is generally a
very slow time for the server but still active.

There was nothing abnormal in the ASE error log. We did
find metedata errors in the Jboss log 30 minutes before the
misbehavior indicating the server could not find an index
descriptor. Just before the problem there was another
application error stating the following.

"Metadata accessor information was not found on this
database. Please install the required tables as mentioned in
the jConnect documentation"

Obviously I am watching our open objects and open index
descriptors like a hawk and I am going to bump them up about
20% in the next day or so.

Usually when we have this we get an application error and we
get a notification in the error log about descriptor
shortages. Again, the ASE error log was clean. We have
never experienced this type of behavior before.

Anybody experience this behavior or have any ideas about the
possibility of server outage caused by metadata descriptor
shortages?


ET Posted on 2010-06-29 21:15:39.0Z
From: ET <dbccjob@gmail.Com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; fr; rv:1.9.1.10) Gecko/20100512 Thunderbird/3.0.5
MIME-Version: 1.0
Newsgroups: sybase.public.ase.general
Subject: Re: Problems after Metadata descripor error
References: <4c2a430d.384e.1681692777@sybase.com>
In-Reply-To: <4c2a430d.384e.1681692777@sybase.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4c2a627b$1@forums-1-dub>
Date: 29 Jun 2010 14:15:39 -0700
X-Trace: forums-1-dub 1277846139 10.22.241.152 (29 Jun 2010 14:15:39 -0700)
X-Original-Trace: 29 Jun 2010 14:15:39 -0700, vip152.sybase.com
Lines: 61
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29328
Article PK: 78560

Hi

>>>>>>>>> Server started refusing application logins.
Did you check if your number of max user connections was not reached ?
Sounds like a connection pool allocating too many connections because
one massive query is making everything slow (i/o, lock) and suddenly
everything is piling up ...
But if there is no error in the errorlog ...

Generally these metadata errors are related to java apps using metadata
and not finding the jdbc metadata stored procs. Either these procs are
not there (but you would see the error since a long time in jboss) or
ASE was so heavily loaded that trying to execute sp_mda times out for
instance...

I do not think increasing the open objects/index limit will solve the
problem. You should try to identify the SELECT that was killed.


Emmanuel

Le 29/06/2010 21:01, jbuhl a écrit :

> ASE 15.0.3 ESD #2
> Solaris 10
>
> Server started refusing application logins. DBA logs in via
> isql and gets connected fine but when he issues "use
> database.." server disconnects his session without message.
> DBA logs back in and immediately issues shutdown command and
> gets message waiting for process to complete. He then logs
> in via dbartisan gui admin tool. This tool pops serveral
> error screens (did not record) and then settles in and he is
> able to see a list of processes; one of which is running a
> select operation. He kills this process via the GUI admin
> tool and the server shuts down clean. He restarts ASE
> server and everything has been fine for a couple days.
> This happened on a Sunday afternoon which is generally a
> very slow time for the server but still active.
>
> There was nothing abnormal in the ASE error log. We did
> find metedata errors in the Jboss log 30 minutes before the
> misbehavior indicating the server could not find an index
> descriptor. Just before the problem there was another
> application error stating the following.
>
> "Metadata accessor information was not found on this
> database. Please install the required tables as mentioned in
> the jConnect documentation"
>
> Obviously I am watching our open objects and open index
> descriptors like a hawk and I am going to bump them up about
> 20% in the next day or so.
>
> Usually when we have this we get an application error and we
> get a notification in the error log about descriptor
> shortages. Again, the ASE error log was clean. We have
> never experienced this type of behavior before.
>
> Anybody experience this behavior or have any ideas about the
> possibility of server outage caused by metadata descriptor
> shortages?


Manish Negandhi [TeamSybase] Posted on 2010-06-30 09:20:55.0Z
From: "Manish Negandhi [TeamSybase]" <no_spam_negandhi.manish@gmail.com>
Newsgroups: sybase.public.ase.general
References: <4c2a430d.384e.1681692777@sybase.com>
Subject: Re: Problems after Metadata descripor error
Lines: 43
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4c2b0c77@forums-1-dub>
Date: 30 Jun 2010 02:20:55 -0700
X-Trace: forums-1-dub 1277889655 10.22.241.152 (30 Jun 2010 02:20:55 -0700)
X-Original-Trace: 30 Jun 2010 02:20:55 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29329
Article PK: 78562


<jbuhl> wrote in message news:4c2a430d.384e.1681692777@sybase.com...
> ASE 15.0.3 ESD #2
> Solaris 10

> "Metadata accessor information was not found on this
> database. Please install the required tables as mentioned in
> the jConnect documentation"

You need to install metadata Stored Procedures used by the JDBC drivers. The
procs should be available in
$SYBASE/jConnect-*_0/sp/sql_server.sql. You might want to test same Java
apps in Test/Dev env to see if you get similar problems.

> ASE 15.0.3 ESD #2
> Solaris 10
>
> Server started refusing application logins. DBA logs in via
> isql and gets connected fine but when he issues "use
> database.." server disconnects his session without message.
> DBA logs back in and immediately issues shutdown command and
> gets message waiting for process to complete. He then logs
> in via dbartisan gui admin tool. This tool pops serveral
> error screens (did not record) and then settles in and he is
> able to see a list of processes; one of which is running a
> select operation. He kills this process via the GUI admin
> tool and the server shuts down clean. He restarts ASE
> server and everything has been fine for a couple days.
> This happened on a Sunday afternoon which is generally a
> very slow time for the server but still active.
>

It will be a good idea to capture any errors during such scenarios such as
capture sqltext, showplan and sysprocesses information to see what is
happening


-HTH
Manish Negandhi
[TeamSybase]


Manish Negandhi [TeamSybase] Posted on 2010-06-30 09:20:56.0Z
From: "Manish Negandhi [TeamSybase]" <no_spam_negandhi.manish@gmail.com>
Newsgroups: sybase.public.ase.general
References: <4c2a430d.384e.1681692777@sybase.com>
Subject: Re: Problems after Metadata descripor error
Lines: 44
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5931
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5931
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4c2b0c78@forums-1-dub>
Date: 30 Jun 2010 02:20:56 -0700
X-Trace: forums-1-dub 1277889656 10.22.241.152 (30 Jun 2010 02:20:56 -0700)
X-Original-Trace: 30 Jun 2010 02:20:56 -0700, vip152.sybase.com
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29330
Article PK: 78563


<jbuhl> wrote in message news:4c2a430d.384e.1681692777@sybase.com...
> ASE 15.0.3 ESD #2
> Solaris 10

> "Metadata accessor information was not found on this
> database. Please install the required tables as mentioned in
> the jConnect documentation"

You need to install metadata Stored Procedures used by the JDBC drivers. The
procs should be available in
$SYBASE/jConnect-*_0/sp/sql_server.sql. You might want to test same Java
apps in Test/Dev env to see if you get similar problems.

> ASE 15.0.3 ESD #2
> Solaris 10
>
> Server started refusing application logins. DBA logs in via
> isql and gets connected fine but when he issues "use
> database.." server disconnects his session without message.
> DBA logs back in and immediately issues shutdown command and
> gets message waiting for process to complete. He then logs
> in via dbartisan gui admin tool. This tool pops serveral
> error screens (did not record) and then settles in and he is
> able to see a list of processes; one of which is running a
> select operation. He kills this process via the GUI admin
> tool and the server shuts down clean. He restarts ASE
> server and everything has been fine for a couple days.
> This happened on a Sunday afternoon which is generally a
> very slow time for the server but still active.
>

It will be a good idea to capture any errors during such scenarios such as
capture sqltext, showplan and sysprocesses information to see what is
happening


-HTH
Manish Negandhi
[TeamSybase]


jbuhl Posted on 2010-07-07 17:35:58.0Z
Sender: 43d4.4c34b935.1804289383@sybase.com
From: jbuhl
Newsgroups: sybase.public.ase.general
Subject: Re: Problems after Metadata descripor error
X-Mailer: WebNews to Mail Gateway v1.1t
Message-ID: <4c34bafe.440e.1681692777@sybase.com>
References: <4c2b0c78@forums-1-dub>
NNTP-Posting-Host: 10.22.241.41
X-Original-NNTP-Posting-Host: 10.22.241.41
Date: 7 Jul 2010 10:35:58 -0700
X-Trace: forums-1-dub 1278524158 10.22.241.41 (7 Jul 2010 10:35:58 -0700)
X-Original-Trace: 7 Jul 2010 10:35:58 -0700, 10.22.241.41
Lines: 20
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29336
Article PK: 78568


> <jbuhl> wrote in message
> > news:4c2a430d.384e.1681692777@sybase.com... ASE 15.0.3
> > ESD #2 Solaris 10
>
> > "Metadata accessor information was not found on this
> > database. Please install the required tables as
> > mentioned in the jConnect documentation"
>
> You need to install metadata Stored Procedures used by the
> JDBC drivers. The procs should be available in
> $SYBASE/jConnect-*_0/sp/sql_server.sql. You might want to
> test same Java apps in Test/Dev env to see if you get
> similar problems.

This is not the problem. This system has been in production
for years supporting JDBC clients. After we rebooted the
system everything was fine. It was something to do with
running out of open object descriptors or index descriptors.
instead of reusing object/index descriptors it became
confused which interrupted processing.


ET Posted on 2010-07-07 19:31:18.0Z
From: ET <dbccjob@gmail.Com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; fr; rv:1.9.1.10) Gecko/20100512 Thunderbird/3.0.5
MIME-Version: 1.0
Newsgroups: sybase.public.ase.general
Subject: Re: Problems after Metadata descripor error
References: <4c2b0c78@forums-1-dub> <4c34bafe.440e.1681692777@sybase.com>
In-Reply-To: <4c34bafe.440e.1681692777@sybase.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
NNTP-Posting-Host: vip152.sybase.com
X-Original-NNTP-Posting-Host: vip152.sybase.com
Message-ID: <4c34d606@forums-1-dub>
Date: 7 Jul 2010 12:31:18 -0700
X-Trace: forums-1-dub 1278531078 10.22.241.152 (7 Jul 2010 12:31:18 -0700)
X-Original-Trace: 7 Jul 2010 12:31:18 -0700, vip152.sybase.com
Lines: 34
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29337
Article PK: 78570

You should try to identify the select query your dba killed and made the
shutdown possible.

There is nothing in the errorlog or in jboss log. It can be a bad query
that overloaded the system. You think about metadata descriptor shortage
but there is apparently no proof of that.

Did you try to log a case at sybase support ?

Emmanuel



Le 07/07/2010 19:35, jbuhl a écrit :

>> <jbuhl> wrote in message
>>> news:4c2a430d.384e.1681692777@sybase.com... ASE 15.0.3
>>> ESD #2 Solaris 10
>>
>>> "Metadata accessor information was not found on this
>>> database. Please install the required tables as
>>> mentioned in the jConnect documentation"
>>
>> You need to install metadata Stored Procedures used by the
>> JDBC drivers. The procs should be available in
>> $SYBASE/jConnect-*_0/sp/sql_server.sql. You might want to
>> test same Java apps in Test/Dev env to see if you get
>> similar problems.
>
> This is not the problem. This system has been in production
> for years supporting JDBC clients. After we rebooted the
> system everything was fine. It was something to do with
> running out of open object descriptors or index descriptors.
> instead of reusing object/index descriptors it became
> confused which interrupted processing.


"Mark A. Parsons" <iron_horse Posted on 2010-07-07 20:04:30.0Z
From: "Mark A. Parsons" <iron_horse@no_spamola.compuserve.com>
User-Agent: Thunderbird 1.5.0.10 (Windows/20070221)
MIME-Version: 1.0
Newsgroups: sybase.public.ase.general
Subject: Re: Problems after Metadata descripor error
References: <4c2b0c78@forums-1-dub> <4c34bafe.440e.1681692777@sybase.com>
In-Reply-To: <4c34bafe.440e.1681692777@sybase.com>
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: <4c34ddce$1@forums-1-dub>
Date: 7 Jul 2010 13:04:30 -0700
X-Trace: forums-1-dub 1278533070 10.22.241.152 (7 Jul 2010 13:04:30 -0700)
X-Original-Trace: 7 Jul 2010 13:04:30 -0700, vip152.sybase.com
Lines: 45
Path: forums-1-dub!not-for-mail
Xref: forums-1-dub sybase.public.ase.general:29338
Article PK: 78573

It may not have any bearing on your situation but fwiw ... and I'm a little hazy on some of the details ...

-----------

A few years ago I had a client that had just released a new version of an internal program.

Within 24 hours of the release we started getting 'out of object descriptors' errors and the ASE becamse sluggish and
would sometimes freeze up.

We eventually tracked the issue back to a bug in the application which was using prepared statements. While the
objective of prepared statements is to re-use said statements once they've been initialized, this program was
initializing a new prepared statement each time a particular piece of code was hit (and it wasn't releasing the
previously prepared statement).

NOTE: When the dataserver receives the initial statement preparation command it will create (in essence) a temporary
stored proc. Follow-on calls to that prepared statement are then handled by the temporary proc. This temporary stored
proc will use up an object descriptor.

So in this case the constant initialization of new prepared statements caused the dataserver to continually create new
temporary stored procs, which eventually used up all of object descriptors.

Disconnecting the application, or bouncing the dataserver, would obviously (?) force a drop of the temporary procs and
the release of the associated object descriptors.

jbuhl wrote:
>> <jbuhl> wrote in message
>>> news:4c2a430d.384e.1681692777@sybase.com... ASE 15.0.3
>>> ESD #2 Solaris 10
>>> "Metadata accessor information was not found on this
>>> database. Please install the required tables as
>>> mentioned in the jConnect documentation"
>> You need to install metadata Stored Procedures used by the
>> JDBC drivers. The procs should be available in
>> $SYBASE/jConnect-*_0/sp/sql_server.sql. You might want to
>> test same Java apps in Test/Dev env to see if you get
>> similar problems.
>
> This is not the problem. This system has been in production
> for years supporting JDBC clients. After we rebooted the
> system everything was fine. It was something to do with
> running out of open object descriptors or index descriptors.
> instead of reusing object/index descriptors it became
> confused which interrupted processing.