In case the performance of a UCS system is bad, here are some basic tests.
Memory
If the server acts after a reboot fine and speedy again, please see this article: Problem: Performance of UCS system is Bad
Groupcache
-
nss/group/cachefile/check_member
Set the variable nss/group/cachefile/check_member to no. This means that when the group cache is rebuilt, there is no further LDAP query for each group member. -
nss/group/cachefile/invalidate_interval
The variable is currently set to recreate the group cache every 15 minutes. Since the group cache will also be recreated after changes, it is sufficient to do this once a day.
Quota
If you do not currently use a quota, please set the UCR variable quota/userdefault to no.
Performance Guide
Please check the UCS Performance Guide
Further links
A great article about Linux Performance Analysis in 60,000 Milliseconds