| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| thrashed disk |
« View previous topic :: View next topic » |
| Author |
Message
|
| sebulba_dz |
 Posted: Mon Oct 03, 2005 4:05 am Post subject: thrashed disk |
 |
|
Novice
Joined: 02 Jun 2003
Posts: 20
Location: UK
|
hi.
i'm running a mq5.3 cluster on 2 similar solaris machines.
one of my mq boxes runs into high disk utilisation on the mq disk partition occationally. i've done a compare of iostat with the 2 machines and the stats is consistent on both machines except for the read operation, which blows up sporadically though i don't know what's causing it. i run a couple of listener process and no heavy applications in the server..
has anyone encountered something like this before? any ideas how i can pinpoint what's performing the read operation or which file's being read? |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Mon Oct 03, 2005 4:07 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Can you correlate the times of high disk activity to the times of high application activity?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| sebulba_dz |
| Posted: Mon Oct 03, 2005 4:54 am Post subject: |
|
|
Novice
Joined: 02 Jun 2003
Posts: 20
Location: UK
|
there's no correlation at all. the occurance is quite sporadic, and it even happens when there's low traffic.
i found the truss command can be used to find out what system calls a process is performing. i haven't tried it on the said server as it's in production and truss will probably slow down the logging. i did a
| Code: |
| truss -t read,open,close,write -p [i]amqhasmx pid[/i] |
in my UAT, but it seems like amqhasmx is all write. so i guess the problem doesn't lie with the log.. |
|
| Back to top |
|
|
| sebulba_dz |
| Posted: Fri Oct 07, 2005 2:33 am Post subject: |
|
|
Novice
Joined: 02 Jun 2003
Posts: 20
Location: UK
|
just found the cause of this disk I/O problem.
i have a contact admin plugin for hp openview and there were error messages :
| Code: |
The UNIX MQControl Local Manager logfile /var/nsq/log/nsqmq.log size of (54108.00) bytes has exceeded the (50000.00) threshold.
|
running iostat, i cat the logs to /dev/null, stopped the openview agent and saw a dramatic decline in disk I/O. however, when i turn it back on, the I/O goes up again. has anyone using contact admin or openview encountered this problem before?
at this point in time it's probably no longer an mq issue..but my hp openview administrator's enjoying the sun in spain for 2 weeks and i'm facing the grim fact that i'll have to bounce the servers if this io problem continues  , so any insights or past experience with contact admin/openview will be appreciated.. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri Oct 07, 2005 8:24 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Have the openview agent put it's logs somewhere where they will not interfere with mq space (different mount). That should keep your MQ happy even though you might have monitoring problems...
Enjoy  |
|
| 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
|
|
|
|
