Discussion:
Queue Manager Performance Analysis
David Awerbuch (BLOOMBERG/ 120 PARK)
2014-07-16 14:51:54 UTC
Permalink
Hi listers,

We have a queue manager with which we are having issues with the response time for certain PCF requests, and I am hoping someone can help clue me in to things to look at.

note - while these time trials are using runmqsc, we are experiencing very similar completion times using a PCF client running off-host - runmqsc does not appear to be a part of the issue.

Appreciate any assistance anyone can offer at this time. The queue manager is running, but the delayed response times are impacting other functionality that relies on the queue manager (monitoring, remote configuration, etc.)

Thanks,
David


SunOS 5.10 147440-13 sun4u sparc SUNW,SPARC-Enterprise
M4000, 16x2400MHz SPARC64

=== output from 'top' ===
load averages: 25.00, 26.28, 25.81
1362 processes:1349 sleeping, 2 running, 1 stopped, 10 on cpu
CPU states: 30.9% idle, 26.9% user, 42.2% kernel, 0.0% iowait, 0.0% swap
Memory: 64.0G real, 29.6G free, 30.2G swap in use, 68.9G swap free
**** does not appear to be a resource issue, as there is plenty of memory, no swapping, no iowait, available CPU time, not a lot of processes contending for CPU cycles, though the load averages appear to be a bit high.

WMQ Version 7.0.1.4
2,622 Local Queues
15,965 Queue Handles
849 Channels
10,937 Active channel instances

-- queues --

$ time echo 'dis ql(*)' | runmqsc PRODQMGR| grep QUEUE | wc -l
2622
real 0m1.04s
user 0m0.29s
sys 0m0.69s

$ time echo 'dis qs(*)' | runmqsc PRODQMGR| grep QUEUE | wc -l
2747
real 0m2.31s
user 0m0.62s
sys 0m0.97s

$ time echo 'dis qs(*) type(handle)' | runmqsc PRODQMGR| grep QUEUE | wc -l
15965
real 2m48.48s <<<<<
user 0m1.79s
sys 0m3.36s

-- channels --

$ time echo 'dis chl(*)' | runmqsc PRODQMGR| grep CHANNEL | wc -l
849
real 0m1.00s
user 0m0.28s
sys 0m0.45s

$ time echo 'dis chs(*)' | runmqsc PRODQMGR| grep CHANNEL | wc -l
10937
real 0m34.18s <<<<<
user 0m27.52s
sys 0m4.46s


$ ps -fu mqm | grep amqrmppa | grep PRODQMGR| wc -l
115
$ ps -fu mqm | grep amqzlaa0 | grep PRODQMGR| wc -l
185


$ mqconfig -v 7.0
mqconfig: V3.7 analyzing Solaris 10 (sparc) settings for WebSphere MQ V7.0
mqconfig: You do not have a group.mqm project configured. IBM recommends
that you configure a group.mqm project with resource limits for
WebSphere MQ, but you can run queue managers under other projects.
If you plan to use a different project for WebSphere MQ, rerun
mqconfig with the -p option to analyze that project.
mqconfig: No project given. Analyzing the current project.
Project user.mqm (): System V Semaphores
max-sem-ids 16 of 2048 sets (0%) IBM>=1024 PASS

Project user.mqm (): System V Shared Memory
max-shm-ids 72 of 65536 sets (0%) IBM>=1024 PASS
max-shm-memory 131941395333120 bytes IBM>=4294967296 PASS

Project user.mqm (): Other Settings
max-file-descriptor 10000 descriptors IBM>=10000 PASS

Shell Default Options ()
ksh bgnice:off IBM:off PASS



<< "Once the game is over, the king and the pawn go back into the same box." - Anon >>

To unsubscribe, write to ***@LISTSERV.MEDUNIWIEN.AC.AT and,
in the message body (not the subject), write: SIGNOFF MQSERIES
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html
Tim Zielke
2014-07-16 15:21:46 UTC
Permalink
Hi David,

For those examples you gave where your time command comes back with a high elapsed time, are those timings consistent when you run that same command multiple times, or do the timings fluctuate between acceptable and unacceptable for the same command when run multiple times?

Is this a non-global zone? If so, have you verified if the global zone is CPU constrained?

Thanks,
Tim

From: MQSeries List [mailto:***@LISTSERV.MEDUNIWIEN.AC.AT] On Behalf Of David Awerbuch (BLOOMBERG/ 120 PARK)
Sent: Wednesday, July 16, 2014 9:52 AM
To: ***@LISTSERV.MEDUNIWIEN.AC.AT
Subject: Queue Manager Performance Analysis

Hi listers,

We have a queue manager with which we are having issues with the response time for certain PCF requests, and I am hoping someone can help clue me in to things to look at.

note - while these time trials are using runmqsc, we are experiencing very similar completion times using a PCF client running off-host - runmqsc does not appear to be a part of the issue.

Appreciate any assistance anyone can offer at this time. The queue manager is running, but the delayed response times are impacting other functionality that relies on the queue manager (monitoring, remote configuration, etc.)

Thanks,
David



SunOS 5.10 147440-13 sun4u sparc SUNW,SPARC-Enterprise
M4000, 16x2400MHz SPARC64

=== output from 'top' ===
load averages: 25.00, 26.28, 25.81
1362 processes:1349 sleeping, 2 running, 1 stopped, 10 on cpu
CPU states: 30.9% idle, 26.9% user, 42.2% kernel, 0.0% iowait, 0.0% swap
Memory: 64.0G real, 29.6G free, 30.2G swap in use, 68.9G swap free
**** does not appear to be a resource issue, as there is plenty of memory, no swapping, no iowait, available CPU time, not a lot of processes contending for CPU cycles, though the load averages appear to be a bit high.

WMQ Version 7.0.1.4
2,622 Local Queues
15,965 Queue Handles
849 Channels
10,937 Active channel instances

-- queues --

$ time echo 'dis ql(*)' | runmqsc PRODQMGR| grep QUEUE | wc -l
2622
real 0m1.04s
user 0m0.29s
sys 0m0.69s

$ time echo 'dis qs(*)' | runmqsc PRODQMGR| grep QUEUE | wc -l
2747
real 0m2.31s
user 0m0.62s
sys 0m0.97s

$ time echo 'dis qs(*) type(handle)' | runmqsc PRODQMGR| grep QUEUE | wc -l
15965
real 2m48.48s <<<<<
user 0m1.79s
sys 0m3.36s

-- channels --

$ time echo 'dis chl(*)' | runmqsc PRODQMGR| grep CHANNEL | wc -l
849
real 0m1.00s
user 0m0.28s
sys 0m0.45s

$ time echo 'dis chs(*)' | runmqsc PRODQMGR| grep CHANNEL | wc -l
10937
real 0m34.18s <<<<<
user 0m27.52s
sys 0m4.46s


$ ps -fu mqm | grep amqrmppa | grep PRODQMGR| wc -l
115
$ ps -fu mqm | grep amqzlaa0 | grep PRODQMGR| wc -l
185


$ mqconfig -v 7.0
mqconfig: V3.7 analyzing Solaris 10 (sparc) settings for WebSphere MQ V7.0
mqconfig: You do not have a group.mqm project configured. IBM recommends
that you configure a group.mqm project with resource limits for
WebSphere MQ, but you can run queue managers under other projects.
If you plan to use a different project for WebSphere MQ, rerun
mqconfig with the -p option to analyze that project.
mqconfig: No project given. Analyzing the current project.
Project user.mqm (): System V Semaphores
max-sem-ids 16 of 2048 sets (0%) IBM>=1024 PASS

Project user.mqm (): System V Shared Memory
max-shm-ids 72 of 65536 sets (0%) IBM>=1024 PASS
max-shm-memory 131941395333120 bytes IBM>=4294967296 PASS

Project user.mqm (): Other Settings
max-file-descriptor 10000 descriptors IBM>=10000 PASS

Shell Default Options ()
ksh bgnice:off IBM:off PASS



<< "Once the game is over, the king and the pawn go back into the same box." - Anon >>

________________________________
List Archive<http://listserv.meduniwien.ac.at/archives/mqser-l.html> - Manage Your List Settings<http://listserv.meduniwien.ac.at/cgi-bin/wa?SUBED1=mqser-l&A=1> - Unsubscribe<mailto:***@LISTSERV.MEDUNIWIEN.AC.AT?subject=Unsubscribe&BODY=signoff%20mqseries>

Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com<http://www.lsoft.com/resources/manuals.asp>

To unsubscribe, write to ***@LISTSERV.MEDUNIWIEN.AC.AT and,
in the message body (not the subject), write: SIGNOFF MQSERIES
Instructions for managing your mailing list subscription are provided in
the Listserv General Users Guide available at http://www.lsoft.com
Archive: http://listserv.meduniwien.ac.at/archives/mqser-l.html

Loading...