ovs_rcu block 8000ms or 1600ms
yangzhixiang12 opened this issue · 1 comments
hi；
in our environment, ovs-vswitchd occasionally ovs-rcu block 8000ms or 1600ms problems，Usually occurs during node restart or pod upgrade process on the node；Print the following log：This problem causes the main thread to be busy and unable to add ovs port；Cause vm or pod creation to fail；
Is there any relevant patch that can solve or relieve this problem?
Looks like something keeps the main thread busy for 8 seconds in the log snippet above. Could be genuine operations that take a long time. But could also be the kernel being too busy. Maybe you want to take a look at the following blog and follow it for troubleshooting: https://developers.redhat.com/articles/2024/01/31/debugging-open-vswitch-logs-long-poll-interval-blocked-waiting