Fwd: HELP...Many DB2 Ctlg Deadlocks w/multiple BINDs...

Charles Valentin

Fwd: HELP...Many DB2 Ctlg Deadlocks w/multiple BINDs...


---------------------------------------------------------------------------------
Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at [login to unmask email] Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm
Attachments

  • import1 (3.6k)

Michael Ebert

Re: Fwd: HELP...Many DB2 Ctlg Deadlocks w/multiple BINDs...
(in response to Charles Valentin)
He's at home, enjoying the weekend with his family, of course!

Actually there's not much I can tell you. Generally our BIND activity
isn't that high that we have any problems. During migrations or after
application of a significant number of PTFs, I've usually run mass-rebinds
(and Cat/Dir REORGs - about every 9 months). It used to be the case that
it didn't pay to run the mass-rebinds in parallel because of lock
contention. However, when migrating our test & prod DB2s to V7, I've run
two in parallel without problems.
I've taken care that if you list the packages in a COLLID.PACKAGE order,
one job had the first half and the other the second half. Since this is
the clustering order, the jobs should be accessing widely separated pages
and no contention should occur (and it didn't).

The best tip to reduce SPT01 (and SYSPKAGE) size is to reduce the number
of packages. We used to keep two or three versions of every package; but
I've managed to convince people that there is no point in keeping an older
version after the new one is in operation for more than a month - any
problem discovered then would be solved by a new version, not fallback to
an old one. Actually I don't think we've ever done a fallback. Maybe you
can use a similar approach. I've also specified a low freespace (FREEPAGE
0 PCTFREE 5) for SYSPKAGE (and thus SPT01) which doesn't cause any
problems in our environment.

Back to the easy chair now...

Dr. Michael Ebert
DB2 Database Administrator
aMaDEUS Data Processing
Erding / Munich, Germany




To:
[login to unmask email]
Subject:
HELP...Many DB2 Ctlg Deadlocks w/multiple BINDs...
HeLP.........
Any recommendations for reducing "deadlocks" in SYSDBASE/SYSPKAGE during
execution of many concurrent BINDs in busy 2-way Data Sharing environment?
Our DB2 Test/Devl system is getting "hammered" by a significant increase
in
Application Pgmrs & DBA's all submitting Batch Compile/Bind jobs on both
L-Pars of a Two-Way Data Sharing environment.
We're at DB2 v7.1 (RSU-0309 + Hipers) and z/OS 1.4 at RSU-0303
maintenance.
Our IRLMRWT=60 and DEADLOCK time =5.
We're REORGing DB2 Catlg/Dir this weekend(last done approx. 2 yrs ago).
Besides REORG, is there anything else we can do to alleviate this problem.
I've opened pmr with IBM Support and FTP'ed DB2/SYSlog documentation.
Ahhhhhh, Where's Dr. Ebert when I need him?
DSNDB01.SPT01 is at (4) 3390's (any tips on reducing SPT01 during Reorg?)
Please, ANY recommendations, hints, tips will be "greatly" appreciated !!!

Thank You,
CharlesV.
DB2 Systems Support


---------------------------------------------------------------------------------
Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at [login to unmask email] Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm