Is the admin console meant to be very slow for operations? Is there a way that we can use it to better perform like some configuration or something. Please advice.
It was not made clear in my prior post but checking S_QOS_DATA was just an example. That and the CM_ are the ones that typically will have high fragmentation leading to performance problems. There are a lot of CM_ tables and which ones are most likely to effect performance varies as per the installation but some common ones are CM_COMPUTER_SYSTEM, CM_COMPUTER_SYSTEM_ATTR, CM_DEVICE, CM_DEVICE_ATTRIBUTE.
If this is UIM 8.2 or 8.3, reconfigure the custom groups from Dynamic to Static.
I am facing the same issue in 8.1.
To add on to what David said, increasing the memory of the service_host probe might improve performance.
In IM, open the raw configure menu (Ctrl + Right Click) and check on the java settings.
The default is:
You can try changing this to the following:
Make sure that the server is able to adequately provide the memory you configure.
Additionally check the database tables for fragmentation.
In MS SQL go to the Management Studio > expand the tree of tables for the UIM database > scroll down to S_QOS_DATA > expand the tree to show the Indexes > right click on an index and select Rebuild. The popup 'Rebuild Indexes' window will report the percentage of fragmentation. Just click OK to run the defrag.
Changing the group type to static from dynamic and also adding more java memory did not help. I am yet to work on the DB defrag and hoping that should resolve this. Thanks a lot for all the advice.
Thanks a lot David.
I did check on the DB and yes, we have a huge performance issue to address.
On the other hand, increasing the java memory in service_host probe did not make any difference. I was wondering if increasing the same in ppm will make any difference. Any comments please?
I don't have anything on increasing the memory for java for either of them. It could be the increase in memory was not enough and doubling it might provide the improvement. Of course this is dependent upon plenty of free system physical RAM. Typically the memory increase is done for probes that have a lot of information to work with such as the vmware probe, or snmpcollector when there are many profiles.
Retrieving data ...