<<We do have them split out into their own subsystem and class and they run
at priority 90 with a restricted memory pool>>
That's what I do, too.
Paul Nelson
Cell 708-670-6978
Office 409-267-4027
nelsonp@xxxxxxxxxxxxx
-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Mike
Cunningham
Sent: Thursday, July 09, 2015 2:54 PM
To: Midrange Systems Technical Discussion
Subject: RE: Speeding up ODBC
I should have added I bit more on the high CPU issue. I would like to see
every processor on campus running around 80% utilization so I know I was
getting good use of the hardware. Way to many times I see windows servers
that don't go over 10% use for days but repeatedly get told that we need
that server and it can't be combined with tasks on some other server because
that might crash the server.
This particular linux application will sometimes have 10-15 QZDASOINIT jobs
running on the system and many times multiple ones are showing high CPU use.
I looked as I was typing this and there were 12 jobs showing some CPU use on
a WRKACTJOB over a 2 second period. Three of them were showing CPU use of
between 22% and 34% each. That CPU use was gone 2 seconds later. The
majority of the time they are using 1-2% and many times <1% of the CPU. It's
when multiple of them hit high CPU and sustain for a long time that causes
some issues for other users. We do have them split out into their own
subsystem and class and they run at priority 90 with a restricted memory
pool
-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
CRPence
Sent: Thursday, July 09, 2015 3:18 PM
To: midrange-l@xxxxxxxxxxxx
Subject: Re: Speeding up ODBC
On 09-Jul-2015 11:21 -0600, Mike Cunningham wrote:
On Thursday, July 09, 2015 1:16 PM Vernon Hamberg wrote:
Yup - in talking with one of our SQL Server folks, he said that
things like SQL Server will take a statement like that and basically
do a SELECT * FROM TABLE on the remote server - no WHERE clause is
passed along.
Then they do the filtering on the client. Crystal Reports was/is like
that, too - one was encouraged in CR to turn on a setting to force
the whole query to run on the remote server.
The behavior of running SELECT * with no selectivity is a well-known
one, according to Rick, the guy I talked to.
<<SNIP>>
Does anyone have any experience with the IBM ODBC Driver for Linux and
it might also behave like this and ignore the where clause?
The client software would be dropping the WHERE criteria, not the driver.
We don't have an issue with speed with incoming ODBC requests from
this linux application but they do have a big CPU hit when they run.
Up to 40-50% when running. And building indexes to try and help has
not.
Maximizing CPU utilization is often a good thing, except if the
utilization impacts other work that is prioritized. Optimize for
efficiency; for sure do *not* optimize to minimize CPU utilization.
--
Regards, Chuck
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe, unsubscribe,
or change list options,
visit:
http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at
http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.