| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Resource problem on solaris |
« View previous topic :: View next topic » |
| Author |
Message
|
| GLL0905 |
 Posted: Tue May 25, 2004 11:53 am Post subject: Resource problem on solaris |
 |
|
Apprentice
Joined: 25 Jun 2001
Posts: 40
Location: gauteng, sa
|
Hi all,
I put 2500 messages through a flow that does a lookup for reply details in a local db2 and converts xml to fix format. But wmqi took forever to process the messages. The top command shows the flow using 48% of the cpu(see below). This is very abnornal as normal cpu usage by the flows is never above 1%. What could the problem be. Part of the overhead could be attributed to the fact that i am not deleting the entries in the database after selecting them. But there were only 3000 entries in the database. Surely it can't take that long to go through 3000 entries in the database.
Any hints will be appreciated.
Thanking you in advance.
gll0905
*************************************************************
load averages: 1.00, 0.97, 0.68 21:14:25
279 processes: 277 sleeping, 2 on cpu
CPU states: 49.6% idle, 49.4% user, 0.4% kernel, 0.6% iowait, 0.0% swap
Memory: 2048M real, 31M free, 21M swap in use, 4073M swap free
PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
4981 mqadmin 16 0 0 245M 135M cpu1 25:23 48.53% DataFlowEngine
4204 mqadmin 4 58 0 85M 51M sleep 1:36 0.73% db2sysc
9631 mqadmin 1 58 0 2584K 1984K cpu0 0:00 0.17% top
2966 mqm 39 59 0 53M 21M sleep 0:19 0.07% amqzlaa0_nd
2213 mqm 5 55 1 20M 4112K sleep 0:56 0.05% amqpcsea
3048 mqadmin 4 59 0 71M 41M sleep 0:16 0.04% db2sysc
2195 mqm 67 59 0 50M 19M sleep 1:19 0.03% amqzlaa0_nd
2124 mqm 4 58 0 19M 4768K sleep 0:13 0.03% amqhasmx
2130 mqm 56 59 0 85M 57M sleep 0:49 0.02% amqzlaa0_nd
2623 mqm 39 58 0 46M 12M sleep 0:42 0.02% amqzlaa0_nd
2202 mqm 45 58 0 82M 48M sleep 0:21 0.02% amqzlaa0_nd
2601 mqm 4 56 1 28M 4896K sleep 0:00 0.02% runmqchl_nd
2456 root 12 58 0 3088K 2064K sleep 0:18 0.01% mibiisa
2179 mqm 5 55 1 24M 6736K sleep 0:37 0.01% amqpcsea
4851 mqm 4 58 0 25M 4208K sleep 0:02 0.01% runmqchl_nd |
|
| Back to top |
|
 |
| fschofer |
| Posted: Tue May 25, 2004 12:15 pm Post subject: |
|
|
Knight
Joined: 02 Jul 2001
Posts: 524
Location: Mainz, Germany
|
Hi,
maybe an active debug trace (happened to me several times)
use mqsireporttrace to check your flows and execution groups.
Greetings
Frank |
|
| Back to top |
|
|
| GLL0905 |
| Posted: Tue Jun 01, 2004 2:33 am Post subject: |
|
|
Apprentice
Joined: 25 Jun 2001
Posts: 40
Location: gauteng, sa
|
Since i installed the latest csd the problem has not come up. I hope the problem is fixed but i need to observe the process for some time.
Thanks
Graham |
|
| Back to top |
|
|
| Missam |
| Posted: Tue Jun 01, 2004 6:16 am Post subject: |
|
|
Chevalier
Joined: 16 Oct 2003
Posts: 424
|
2500 messages
xml to fix format
only 3000 entries in the database
there are sufficeint for broker to take an hour to process.
| Quote: |
| The top command shows the flow using 48% of the cpu(see below). This is very abnornal as normal cpu usage by the flows is never above 1%. |
you may not observed it before if broker starts processing messages.it'll defenetely goes more than 1% of cpu usage.
1) if you can tune the database you can see some performance increase.
2)if you have more than one database select per message.you can make one select and store them in localEnvironment.this can improve performance much better
3)if you can avoid database operations all together.you can see the best performance.
4)if you have huge volume of data you can increase number of instances of the flow |
|
| 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
|
|
|
|
