Description of problem: quota-crawl process related glusterfs mount is seen to be there after quota is disabled and it's seen even after a long time. [root@nfs1 ~]# ps -auxww | grep quota Warning: bad syntax, perhaps a bogus '-'? See /usr/share/doc/procps-3.2.8/FAQ root 5005 1.5 0.8 441808 68088 ? Ssl 04:26 1:30 /usr/sbin/glusterfs -s localhost --volfile-id dist-rep --use-readdirp=no --client-pid -100 -l /var/log/glusterfs/quota-crawl.log /tmp/mnt1wtbSy root 24510 0.0 0.0 103252 812 pts/0 R+ 06:02 0:00 grep quota same issue is seen on the different nodes. Version-Release number of selected component (if applicable): glusterfs-3.6.0.24-1.el6rhs.x86_64 How reproducible: seen on this build. Expected results: quota-crawl process should be post quota is disabled and that's also for a long time as this shows a discrepency of keeping the mount with "/tmp", which itself may security concerns Additional info:
upstream patch: http://review.gluster.org/#/c/12952/ fixes this problem
This bug was accidentally moved from POST to MODIFIED via an error in automation, please see mmccune with any questions
Hi, With versioning, the crawl process related issues are fixed. Every time quota is enabled, it creates a new set of xattrs(by suffixing the numbers), and previously created xattrs and the crawl process gets killed then.