Tracking Resource Limit Facility activity - DB2 for OS/390

Sue Janowitz

Tracking Resource Limit Facility activity - DB2 for OS/390
We are at DB2 V5.1 (OS/390), and use Landmark's TMON for DB2 monitor. We
have just implemented the use of the DB2 resource limit facility (RLF) to
kill threads using 3 specific plans that use too much CPU time. We have
problems with these threads infrequently - maybe once every few weeks - but
with enough impact that we want to automatically kill them before things go
too far.

My question is: is there any way I can track the RLF activity so that I
know when a thread has been killed by it? (Even retrospectively, via a
report after the fact.) In my test case (I put in a RLF table entry to kill
a DSNTEP2 thread with my ID over 1/4 second long), the only way I can see
what happened is the SQL code of -905 returned in my batch job. Nothing in
the DB2 address space, System log, TMON/DB2 as far as I can tell (TMON/DB2
gives a NORMAL reason code). The jobstep gets an 8 return code.

TIA for any help.

Sue Janowitz
New England Financial Information Services
501 Boylston Street
Boston, MA 02116
email: [login to unmask email]

Shaun Z Lombard

Re: Tracking Resource Limit Facility activity - DB2 for OS/390
(in response to Sue Janowitz)
Sue,

I have the same problem, the best I could come up with was to define an
exception definition based on amount of CPU. Example below.
This one should tell you when a thread running plan DISTSERV reaches 2 CPU
seconds(our limit).

HOW LIMIT DET HOW LIMIT WAS DETERMINED DRXAPREC
GT 0 COUNT 9(9)
THREAD CPU TIM THREAD CPU TIME DRASCTIM
GT 00:00:02.0000 HH:MM:SS.TTTT
PLAN NAME PLAN NAME
DRHCPLAN
EQ DISTSERV CHAR(08)

Let me know if you find a better way of doing this.

Thanks

Shaun




Sue Janowitz <[login to unmask email]> on 08/10/99 22:16:34

Please respond to DB2 Data Base Discussion List <[login to unmask email]>

To: [login to unmask email]
cc: (bcc: Shaun Z Lombard/CorpServ/qdot/au)
Subject: Tracking Resource Limit Facility activity - DB2 for OS/390




We are at DB2 V5.1 (OS/390), and use Landmark's TMON for DB2 monitor. We
have just implemented the use of the DB2 resource limit facility (RLF) to
kill threads using 3 specific plans that use too much CPU time. We have
problems with these threads infrequently - maybe once every few weeks - but
with enough impact that we want to automatically kill them before things go
too far.

My question is: is there any way I can track the RLF activity so that I
know when a thread has been killed by it? (Even retrospectively, via a
report after the fact.) In my test case (I put in a RLF table entry to kill
a DSNTEP2 thread with my ID over 1/4 second long), the only way I can see
what happened is the SQL code of -905 returned in my batch job. Nothing in
the DB2 address space, System log, TMON/DB2 as far as I can tell (TMON/DB2
gives a NORMAL reason code). The jobstep gets an 8 return code.

TIA for any help.

Sue Janowitz
New England Financial Information Services
501 Boylston Street
Boston, MA 02116
email: [login to unmask email]








*************************************************************
Opinions contained in this e-mail do not necessarily reflect
the opinions of the Queensland Department of Main Roads, or
of Queensland Transport. If you have received this electronic
mail message in error, please immediately notify the sender
and delete the message from your computer.