Skip to Content
0

Can we use sp_shmdumpconfig to collect memory dump for a WARNING scenarios?

Nov 01, 2016 at 10:00 AM

172

avatar image
Former Member

Dears,

I wanted to know, if we could configure sp_shmdumpconfig to collect memory dump for a WARNING message that hits my ASE Server at random times?

Regards,

Ajith Prabhakaran

10 |10000 characters needed characters left characters exceeded
* Please Login or Register to Answer, Follow or Comment.

4 Answers

Best Answer
Bret Halford
Nov 03, 2016 at 01:28 PM
0

Hi Aj,

The message number for this warning is 10049.

In general, the internal timer depends on receiving regular SIGALRM signals from the OS
so the cause is typically something keeping the OS very busy.
Are there any other errors in the ASE errorlog around the time of these warnings?
What is the ASE version?

Share
10 |10000 characters needed characters left characters exceeded
Bret Halford
Nov 02, 2016 at 11:03 PM
0

Depends on the exact message. ASE uses several different methods to output text to the errorlog.
There are "errors" which are output with a header line giving the error number, severity level, and state value.
There are "messages", which actually do have a message number assigned to them internally, but it isn't output to the log
(I do have a feature request filed to have the message number exposed in the log).
There are also various messages that are just directly printed to the log.

If the warning you are asking about is one that has an internal message number, I (or another support engineer)
can look up that number and you can configure a dump on that "message" number. If the warning is of the third
category, just being printed directly to the log, we can't trigger a CSMD on it.

Please post the text of the exact warning you are asking about, and I'll see if it has a message number.

Cheers,
-bret

Share
10 |10000 characters needed characters left characters exceeded
avatar image
Former Member Nov 03, 2016 at 11:01 AM
0

True, The warnings I was mentioning do not carry an message number assigned.

What do I do for them? I wondering what actually made the engine too busy at that point of time. We do not have schedule jobs or activities running at that time. Ideally these errors, shouldn't have occurred.

Below are the grep of errors from the logs:-

00:0001:00000:00000:2016/10/25 22:32:12.52 kernel Warning: The internal timer is not progressing. If this message is generated multiple times, report to Sybase Technical Support and restart the server (alarminterval=-760).

00:0001:00000:00000:2016/10/25 22:44:25.43 kernel Warning: The internal timer is operational and has recovered from previous failure.

00:0001:00000:00000:2016/10/26 00:33:12.26 kernel Warning: The internal timer is not progressing. If this message is generated multiple times, report to Sybase Technical Support and restart the server (alarminterval=-1192).

00:0001:00000:00000:2016/10/26 00:43:12.23 kernel Warning: The internal timer is operational and has recovered from previous failure.

00:0001:00000:00000:2016/10/30 23:23:16.34 kernel Warning: The internal timer is not progressing. If this message is generated multiple times, report to Sybase Technical Support and restart the server (alarminterval=-797).

Regards,

Aj

Share
10 |10000 characters needed characters left characters exceeded
avatar image
Former Member Nov 06, 2016 at 05:33 AM
0

Hello Bret,

The version we have is

Adaptive Server Enterprise/15.7/EBF 24646 SMP SP134 /P/RS6000/AIX 6.1/ase157sp133x/3925/64-bit/FBO.

These errors seen usually at late our after 10.30-11 PM. Ideally the OS is meant to be pretty much free. We do not have dump also scheduled to run at that point of time. I verified with OS Admin. He said the OS was with no load at that time.

I did a server bounce the other day. The error seems to be moved off.

But curious to understand, why could have this happened.

Meanwhile Bret, How is the message number you decoded?

Regards,

Aj

Share
10 |10000 characters needed characters left characters exceeded