Author |
Message
|
rcp_mq |
Posted: Wed Dec 05, 2012 11:08 am Post subject: MAXMQERRORLOGSIZE in MQ7 |
|
|
 Centurion
Joined: 13 Dec 2011 Posts: 133
|
Can MAXMQERRORLOGSIZE be safely used in MQ 7.0.0.X, It is missing in the MQ 7.0.0.X documentation.
I would appreciate some knowledge sharing on the basis of experience as I might want to use it in the production environment. |
|
Back to top |
|
 |
JosephGramig |
Posted: Wed Dec 05, 2012 12:11 pm Post subject: |
|
|
 Grand Master
Joined: 09 Feb 2006 Posts: 1244 Location: Gold Coast of Florida, USA
|
Don't use that...
Here are my favourite stanza additions:
Code: |
# The following lines are added to the default
CHANNELS:
MaxChannels=1000
MaxActiveChannels=1000
AdoptNewMCACheck=ALL
AdoptNewMCA=ALL
AdoptNewMCATimeout=60
TCP:
KeepAlive=Yes
ListenerBacklog=256
QMErrorLog:
ErrorLogSize=1048576
|
The last two are relevant to your question. Also, if this is not Windows, change TripleWrite to SingleWrite |
|
Back to top |
|
 |
exerk |
Posted: Wed Dec 05, 2012 2:39 pm Post subject: |
|
|
 Jedi Council
Joined: 02 Nov 2006 Posts: 6339
|
JosephGramig wrote: |
...Also, if this is not Windows, change TripleWrite to SingleWrite |
I add the caution that you should verify your storage medium provides full write integrity before doing so. _________________ It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
Back to top |
|
 |
PeterPotkay |
Posted: Wed Dec 05, 2012 3:07 pm Post subject: |
|
|
 Poobah
Joined: 15 May 2001 Posts: 7722
|
Andy Hickson from Hursley posted the following on the MQ List Server on 10-31-2012 during a discussion that touched upon Single versus Triple write.
Quote: |
A surprising number of customers set SingleWrite without fully understanding either the integrity or the performance implications.
Firstly, the integrity implications. Imagine a 4KB page has been written to the recovery log containing all 'A' characters, and then a subsequent write request writes a page of 'B' characters to the same location. If there are ANY circumstances (e.g power failure, HA failover, ...) where it is possible for the resulting data on disk not to be either all A's or all B's then it is NOT safe to set the SingleWrite option.
Secondly the performance implications, this option only has an effect when the queue manager is writing a partial page to the recovery log. The queue manager tries not to write partial pages, and as long as there's sufficient concurrency in the workload the percentage of partial pages written is very small, and therefore the effect of the SingleWrite/TripleWrite decision is very minor.
Take for example a single application writing 256 byte messages and committing after every message. In this case nearly every page will be a partial page and the effect of the SingleWrite/TripleWrite decision will be significant. On the other hand if 50 instances of the same application were running concurrently then hardly any partial pages would be written and the effect of SingleWrite/TripleWrite will be minor.
In my experience, If you don't have access to a very strong understanding of the low level workings of your disk subsystems you should take the safe option of TripleWrite.
|
_________________ Peter Potkay
Keep Calm and MQ On |
|
Back to top |
|
 |
rcp_mq |
Posted: Wed Dec 05, 2012 8:55 pm Post subject: |
|
|
 Centurion
Joined: 13 Dec 2011 Posts: 133
|
I too have never suggested(used) single write. It is analogous to the persistent and impersistent message concept.
@joseph i will try your method.
@Peter thanks for sharing. I could use it as an argument. |
|
Back to top |
|
 |
rcp_mq |
Posted: Thu Dec 06, 2012 7:38 am Post subject: |
|
|
 Centurion
Joined: 13 Dec 2011 Posts: 133
|
Peter, your amazing MQSeries secrets didn't quite appeal to me! |
|
Back to top |
|
 |
Vitor |
Posted: Thu Dec 06, 2012 8:36 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
rcp_mq wrote: |
Peter, your amazing MQSeries secrets didn't quite appeal to me! |
They should. Deep, deep truth is found there. _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
bruce2359 |
Posted: Thu Dec 06, 2012 9:23 am Post subject: |
|
|
 Poobah
Joined: 05 Jan 2008 Posts: 9469 Location: US: west coast, almost. Otherwise, enroute.
|
Vitor wrote: |
rcp_mq wrote: |
Peter, your amazing MQSeries secrets didn't quite appeal to me! |
They should. Deep, deep truth is found there. |
Deep, abiding truth does not come to you; rather, you must search for it. _________________ 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 |
|
 |
Vitor |
Posted: Thu Dec 06, 2012 9:47 am Post subject: |
|
|
 Grand High Poobah
Joined: 11 Nov 2005 Posts: 26093 Location: Texas, USA
|
bruce2359 wrote: |
Vitor wrote: |
rcp_mq wrote: |
Peter, your amazing MQSeries secrets didn't quite appeal to me! |
They should. Deep, deep truth is found there. |
Deep, abiding truth does not come to you; rather, you must search for it. |
But that's one of the links.....  _________________ Honesty is the best policy.
Insanity is the best defence. |
|
Back to top |
|
 |
JosephGramig |
Posted: Thu Dec 06, 2012 11:28 am Post subject: |
|
|
 Grand Master
Joined: 09 Feb 2006 Posts: 1244 Location: Gold Coast of Florida, USA
|
Vitor, you are probably right and I'm "assuming" all SAN used with UNIX will have "full write integrity". I've asked many times only to receive the same blank stare every time. I do this more to get folks that think it will get them some sort of performance gain off my @$$ (even though I know it will never make a difference). The cost of arguing over it is just too high. Never had it fail in 14 years (although I've only changed it in the last few years). |
|
Back to top |
|
 |
rcp_mq |
Posted: Fri Dec 07, 2012 7:26 am Post subject: |
|
|
 Centurion
Joined: 13 Dec 2011 Posts: 133
|
@Vitor
@bruce factually that's both the enticing links below Peter's signature i'm talking about...perhaps i should've authored my statement differently.
(i do carry my gratitude to all the knowledge sharing done by you all...if i say more i'll be jumping out the fence of MQSeries) |
|
Back to top |
|
 |
|