| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Performance Cost of Enabling PERFMEV |
« View previous topic :: View next topic » |
| Author |
Message
|
| jmmcdowell |
 Posted: Mon Jan 05, 2009 1:44 pm Post subject: Performance Cost of Enabling PERFMEV |
 |
|
Newbie
Joined: 05 Jan 2009
Posts: 6
|
Hello fellow MQ Admins--
I am new to the community but looking forward to share experiences and recommendations in the weeks and months that follows.
I have been tasked at my place of employment with setting up Tivoli Omegamon for monitoring our MQ environment. Looking through the product documentation, it looks like Omegamon requires PERFMEV to be enabled in order to do a lots of different types of reporting/alerting. This includes the ability to display information about "highest queue depth" over a period, the "oldest message on a queue", and much more.
We have never investigated turning on PERFMEV on our queue managers in the past, as this option is disabled by default. My question is this:
- Is there a large performance hit (or any at all) to enabling PERFMEV on qmgrs? We have some fairly busy queue managers in our environment and losing performance is really not an option.
- Are most other MQ users out there turning this option on, or leaving it off?
I look forward to your thoughts, Thanks |
|
| Back to top |
|
 |
| PeterPotkay |
| Posted: Mon Jan 05, 2009 2:30 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
We have it turned on across all QMs in our environment. QPASA relies on it.
Since we have always used it, I can't comment on the performance impact. I wouldn't consider us an extreme MQ message environment though, so I doubt it would be an issue for us.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Mon Jan 05, 2009 4:13 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9471
Location: US: west coast, almost. Otherwise, enroute.
|
What platform? What WMQ version?
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| jmmcdowell |
| Posted: Tue Jan 06, 2009 6:56 am Post subject: |
|
|
Newbie
Joined: 05 Jan 2009
Posts: 6
|
We will be running Tivoli on a windows box..
The queue managers themselves run on a variety of platforms, but mostly AIX.. We also have some HPUX, Linux, VMS, and a few z/OS.
Another question about PERFMEV...
Does a queue manager require a cycle after this is enabled? |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Tue Jan 06, 2009 10:02 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
| jmmcdowell wrote: |
Another question about PERFMEV...
Does a queue manager require a cycle after this is enabled? |
I'm pretty sure no, but easy enough to test.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Jan 06, 2009 10:17 am Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9471
Location: US: west coast, almost. Otherwise, enroute.
|
A definite no.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
| jmmcdowell |
| Posted: Tue Jan 06, 2009 11:26 am Post subject: |
|
|
Newbie
Joined: 05 Jan 2009
Posts: 6
|
| bruce2359 wrote: |
| A definite no. |
A definite no to what, Bruce?
No, there is no need to recycle the qmgr?
Or No, there is definately not a significant performance hit? |
|
| Back to top |
|
|
| bruce2359 |
| Posted: Tue Jan 06, 2009 12:49 pm Post subject: |
|
|
Poobah
Joined: 05 Jan 2008
Posts: 9471
Location: US: west coast, almost. Otherwise, enroute.
|
| Quote: |
A definite no to what, Bruce?
No, there is no need to recycle the qmgr?
Or No, there is definately not a significant performance hit? |
Sorry.
There is no need to cycle the qmgr to enable performance events.
Yes, there is/are additional processor cycles and i/o involved in the qmgr capturing performance data. Notice that I did not use the term overhead. If your business requires monitoring (for chargeback, monitoring tools, automation, problem-determination, whatever), then enabling performance events is a business requirement.
The issue (for management) is whether your current hardware is robust enough to continue to meet SLAs. If not, the business requirement for performance monitoring must include a hardware upgrade.
I don't recall a white-paper addressing percent increase in resource utilization (overhead) due to enabling any, some, all, performance event monitoring. I will defer to IBM internals folks.
_________________
I like deadlines. I like to wave as they pass by.
ב''ה
Lex Orandi, Lex Credendi, Lex Vivendi. As we Worship, So we Believe, So we Live. |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
