Sorry, I just mentioned MGTC as an example, IIRC that used to cause
problems with SQL commits in QSQSRVR jobs or something like that.
To avoid the MGTCOL locks you'll need to ENDPFRCOL before the save and
STRPFRCOL after. This will end the QYPSPFRCOL job, which is locking
the *MGTCOL object.
Regards,
Scott Ingvaldson
Senior IBM Support Specialist
Midwest Region Data Center
Fiserv.
-----Original Message-----
From: rob@xxxxxxxxx [mailto:rob@xxxxxxxxx]
Sent: Friday, April 24, 2009 10:20 AM
To: Midrange Systems Technical Discussion
Subject: RE: Save While Active Question
Lot of good stuff in there.
From the backup joblog:
04/24/09 05:02:49.154072 ENDTCPSVR SERVER(*MGTC)
04/24/09 05:02:49.618967 ENDJOB started for job
156694/QYPSJSVR/QYPSJSVR.
04/24/09 05:02:49.626270 MGTC server ended.
04/24/09 05:10:57.270816 SAVLIB LIB(QMPGDATA) SAVACT(*SYSDFN) ...
04/24/09 05:13:01.693377 Cannot use MGTCOL Q114000024 in QMPGDATA.
04/24/09 06:16:59.903990 STRTCPSVR SERVER(*MGTC)
From the Management Central joblog:
Job 156694/QYPSJSVR/QYPSJSVR ended on 04/24/09 at 05:02:58
*MGTCOL's are an abomination of an object that IBM either needs to fix
or abandon. I submitted a DCR that they abandon them. Obviously they
have locking issues. They also routinely get damaged.
Rob Berendt
--
Group Dekko Services, LLC
Dept 01.073
Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From:
"Ingvaldson, Scott" <scott.ingvaldson@xxxxxxxxxx>
To:
"Midrange Systems Technical Discussion" <midrange-l@xxxxxxxxxxxx>
Date:
04/24/2009 11:04 AM
Subject:
RE: Save While Active Question
Sent by:
midrange-l-bounces@xxxxxxxxxxxx
Save While Active is a wonderful thing. It does not, however, negate
reading the joblog afterwards. Of course there is a TAATOOL, CHKSAVRST,
that is very helpful in analyzing backup joblogs.
It is not a requirement to save every single object on the system to
have a perfectly valid recovery tape. Only the SysAdmin can decide
this, but the MGTCOL object that you point out not being saved here will
be all but useless in a DR situation when you restore to a different
machine. If that object is missing in a recovery secenario it will
automatically be recreated. I would consider omitting the performance
data library from some backups entirely, unless you have a true business
need for this data.
Instead of using SAVACT *SYSDFN try using SAVACT *SYNCLIB and setting
the Save active wait time to a small number like 10 seconds. Run a test
backup and see if anything gets missed. Then you'll know that you have
maintained a consistent state between your objects.
Other things that you might consider when writing a SWA backup is ending
just the items that cause problems and leaving the rest running. For
example, I usually put an ENDTCPSVR SERVER(*MGTC) before a SWA backup
and a STRTCPSVR SERVER(*MGTC) after. That saves you the time of
checking out bogus "not saved" messages every day.
Regards,
Scott Ingvaldson
Senior IBM Support Specialist
Midwest Region Data Center
Fiserv.
-----Original Message-----
From: rob@xxxxxxxxx [mailto:rob@xxxxxxxxx]
Sent: Friday, April 24, 2009 7:28 AM
To: Midrange Systems Technical Discussion
Subject: Re: Save While Active Question
Do not assume that a Save While Active will save everything.
SAVLIB LIB(QMPGDATA) SAVACT(*SYSDFN) ...
CPF3761
Cannot use MGTCOL Q114000024 in QMPGDATA.
MGTCOL Q114000024 is locked by another job
CPF3771 429 objects saved from QMPGDATA. 1 not saved.
I did some checking on a far busier system and didn't see any hits in
the joblog for 'not saved' instead of some stream file data I need to
review.
Rob Berendt
--
Group Dekko Services, LLC
Dept 01.073
Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
midrange.com
