Gluster users - Sep 2017 - Gluster high inode usage on root EC2 volume

Hello,


I have one Gluster server in a cluster of four that serves a single volume.
Three out of the four servers have fine inode usages - hovering around 11-12%,
but one server is using every inode available causing us to get no space left on
disk.


I've listed out the files, there were a ton being used in
/var/lib/misc/glusterfsd/farmcommand/... related to CHANGELOGs which I assuemd
was related to geo-replication, which we have shut off for some time. The last
modified times were near when we shut it down, so I removed them all. However,
this and removing linux src headers only brought me down to 95% inode usage.
Where could these other inodes be being used?
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20170930/ee9e454d/attachment.html>