Need your suggestions on the following issue:
We have a veritas cluster configured on our File Servers with a Replication DR site. We patched up the Nodes :
A lot of our users have been reporting about slowness in accessing these file shares on the cluster. This is an intermittent issue as when we failover to the other node, issue is resolved. After another 5 - 10 days users w ill report the same issue.
We have got the vx explorer logs analysed from symantec support, but nothing seems to indicate the cause of the issue. Support also suggested to collect perfmon from affected users.. but no conclusion. We think, it may be a small tweek in the performance tuning of the product.
Kindly guide to resolve this issue at the earliest.
I've attached the system performance logs from the active node.
Appreciate all the help.
With the time that you describe that the issue happens it sounds like a resource leak. Maybe memory or threads. I would recommend that you watch the server to see how much memory is being used and what processes are using it. Also monitor the number of threads being used and what is using them.
From there I would try to narrow down what you can do to fix the performance issue without a full reboot. You could try force stopping the VCS services then restart VCS to see if performance returns to normal. From there you could try restarting other services/applications to see if they have any impact on performance.
Let us know how you are doing on resolving this issue.
Symantec Support is also available to assist you more directly. If you need more direct help, please open a case with Techncial Support.
1. you can check if fragment is serious, and do defragment regularly.
2. check if maxclient is set properly.
Performance issue is complex, many factors involve.
You can check when it's slow, bottleneck is where?
1. io slow?
2. memory poor?
check one by one,and solve one by one.