On Fri, 1 Feb 2008 09:31:02 -0700, Lance Kopplin <[EMAIL PROTECTED]> wrote:

>
>Given the SMF spikes I've seen when somebody makes a booboo, I would want a
>lockdown method of getting data out of the SMF logstream now [2].
>

Earlier this week I was called by one of my coworkers because SMF offloads
could not keep up on one of our LPARs.  I temporarily moved them from
a "STCHI" service class (importance 2) to SYSSTC while I looked at it.  
We knew from the dumps it was DB2 records (102).   The last time I had
run into something like this (it seems to happen once or twice a year) it
was some sort of expired MQ message causing a problem with generating
a transaction in a CICS region over and over and over and over again.
This time it was worse. 

I had trouble telling what application was causing the problem so I had
my coworker get the DB2 folks involved.  It turned out they turned on some
sort of performance trace in that DB2 subsystem for IBM (it would have been
nice if they monitored the impact or let the "mvs guys" know about it).

Anyway, before all was said and done there were about 150 virtual 
3490E (800M without compression) offloads written.   

So yes... I agree.  The current logger offload program would have been
a huge problem in this situation since it can't delete the records after they
are offloaded.   Our LOGR SMS pool would have run out of space long
before this problem was resolved  (it of course would be larger than
it is today to allow for SMF data and a good fudge factor for spikes). 

Mark
--
Mark Zelden
Sr. Software and Systems Architect - z/OS Team Lead
Zurich North America / Farmers Insurance Group - ZFUS G-ITO
mailto:[EMAIL PROTECTED]
z/OS Systems Programming expert at http://expertanswercenter.techtarget.com/
Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to