RE: Inability to reliably monitor long-running parallel queries

From: Tornblad, John <JTornblad_at_emdeon.com>
Date: Tue, 13 Nov 2012 16:10:19 -0600
Message-ID: <0EC8AF65BC8E9A4E9C4EB86F1882C96E052E1349_at_ZBNAAEEX052.na.webmd.net>



Grin, Mark.

FYI for future readers of this thread... we are requesting an "enhancement request" for SQL Monitoring to *somehow* be able to continue monitoring parallel queries without marking them as "DONE (ERROR)" even though they are still active.

MMON's overly aggressive attack (IMHO) on sessions (parallel servers in this case) which are idle for more than 30 minutes without regard to other concurrent parallel servers still working on the same query under the same QC seems Just Plain Wrong. To recap: SQL Monitoring for a parallel query may or may not work past 30 minutes, even if all your _sqlmon_* parameters are great, just depends on amount / distribution of work on the parallel servers... if any of them sit around for more than 30 minutes, kaboom as far as monitoring is concerned.

Grass roots movement alert! Please chime in to Oracle that you are in favor of an "enhancement" that allows SQL Monitor to continue monitoring your long running parallel queries if you have noticed this problem yourself. We've been told that this "enhancement request" will get attention only if more customers also proclaim "The software, it no worky!" (google this phrase for an image of my favorite Dilbert strip ever). And I suppose we're the first since no comment was made indicating any other customer had already complained. But, we did pursue this for 5+ months until getting to this point... maybe we just get a point for perseverance.

-john

| -----Original Message-----
| From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-
| bounce_at_freelists.org] On Behalf Of Mark W. Farnham
| Sent: Friday, November 02, 2012 2:48 PM
| To: landstander668_at_gmail.com; oracle-l_at_freelists.org
| Subject: RE: Inability to reliably monitor long-running parallel
queries
|
| Kafka would be proud. Well, okay, he wouldn't actually be proud, but
he
| might experience some intellectual amusement that reporting "done with
| error" for a query that is still running and which may in fact
complete
| without error is defined as "not a bug." Damn, I don't see a tuxedo
| disguising this one as a feature!
|
| mwf

</pre>This message is confidential, intended only for the named recipient(s) and may contain information that is privileged or exempt from disclosure under applicable law. If you are not the intended recipient(s), you are notified that the dissemination, distribution, or copying of this message is strictly prohibited. If you receive this message in error or are not the named recipient(s), please notify the sender by return email and delete this message. Thank you.

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Nov 13 2012 - 23:10:19 CET

Original text of this message