Exalogic Virtual: High CPU Usage and IO Wait On Guest vServer, Kworker CPU Usage Is Always High
(Doc ID 2751262.1)
Last updated on FEBRUARY 09, 2021
Applies to:Oracle Exalogic Elastic Cloud Software - Version 220.127.116.11.201020 and later
Information in this document applies to any platform.
On exalogic virtual 18.104.22.168.x, the following is seen on a specific guest vServer, the same is not seen on the other guest vServers at the same time:
1. The %wa is very high, and kworker process CPU usage has been always high:
2. The situation will continue till the vServer is rebooted.
3. The ZFSSA performance is good at the time of the issue, and no issue for other guest vServers on the same exalogic rack.
4. The issue happened radomly on different vServers on the same exalogic rack time to time, one vServer at a time.
5. NFS v4 is used on the NFS mount points on the guest vServer which experiecned this issue.
6. From the packets captrued between the vServer and ZFSSA at the time of the issue, sometimes we can see it's flooded of NFS4ERR_OLD_STATEID packets, sometimes it's flooded of NFS4ERR_STALE_STATEID and NFS4ERR_BAD_STATEID packets.
7. Increase "NFS grace period " setting to 360 from 90 on ZFSSA end, but the issue still happens.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document