[Gluster-users] high number of pending calls
ingard at jotta.no
Fri Sep 1 14:24:22 UTC 2017
We're seeing high? number of pending calls on two of our glusterfs 3.10
We have not tried to tune anything except changing server.event-threads: 2
gluster volume status callpool | grep Pending results in various numbers
but more often that not a fair few of the bricks have 200-400 pending calls.
Is there a way I can debug this further?
The servers are 8x dual 8 core with 64G memory running a distributed
cluster with 4 bricks on each.
The four glusterfsd processes are averaging 100% cpu each and the load
average is around 25-30
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Gluster-users