Thank you for the feedback, Amar!
This is the first time I'm hearing of the Kadalu project. I will keep an eye on that as well as Gluster 10.
Cheers,
On Thu, Sep 9, 2021 at 10:30 AM Amar Tumballi <amar@xxxxxxxxx> wrote:
On Thu, Sep 9, 2021 at 11:13 AM Alan Orth <alan.orth@xxxxxxxxx> wrote:For what it's worth, I disabled quotas and re-enabled them. This of course caused the limits I had set to be erased and GlusterFS to re-run the find/stat on all bricks. Once it finished I was able to add a limit and it was finally accurate.I'm still hoping someone can comment about this. Is it normal, what is the status of quotas in GlusterFS 8.x, is there maintenance I should do, etc?Glad to hear that it worked for you. In general, Quota feature is maintained not actively but based on any critical or CVE bug fixes right now.We needed quota feature for the kadalu project (https://kadalu.io or https://github.com/kadalu/kadalu), for which the crawling quota update turned out to be a major bottleneck. Hence we came up with 'simple-quota' approach which is based on ideas previously discussed around project quotas etc. The feature is already developed and used in kadalu releases, but waiting reviews in glusterfs. Hopefully would make it to glusterfs-10.Thank you,On Thu, Sep 2, 2021 at 1:42 PM Alan Orth <alan.orth@xxxxxxxxx> wrote:Yesterday I noticed that there were still several find / stat processes running on this node (there are two bricks for this volume on this host) after enabling quotas:# ps auxw | grep stat
root 5875 0.0 0.0 112812 976 pts/0 S+ 13:28 0:00 grep --color=auto stat
root 19846 3.9 0.0 121804 2624 ? S Aug31 52:49 /usr/bin/find . -exec /usr/bin/stat {} \ ;
root 19856 3.1 0.0 121784 2536 ? S Aug31 42:02 /usr/bin/find . -exec /usr/bin/stat {} \ ;I waited for them to finish, but my quota information is still incorrect for the two directories I have enabled it on. I ran the quota_fsck.py script on two different hosts and every time it runs it has a different number of objects fixed:# ./quota_fsck.py --sub-dir aorth --fix-issues /mnt/homes-quota-fix /data/glusterfs/sdc/homes
MARKING DIRTY: /data/glusterfs/sdc/homes/aorth
stat on /mnt/homes-quota-fix/aorth
Files verified : 9160
Directories verified : 4670
Objects Fixed : 2920
MARKING DIRTY: /data/glusterfs/sdc/homes/aorth
stat on /mnt/homes-quota-fix/aorth
Files verified : 9160
Directories verified : 4670
Objects Fixed : 3487
MARKING DIRTY: /data/glusterfs/sdc/homes/aorth
stat on /mnt/homes-quota-fix/aorth
Files verified : 9160
Directories verified : 4670
Objects Fixed : 3486What's going on here?Thank you for any help...On Wed, Sep 1, 2021 at 10:41 AM Alan Orth <alan.orth@xxxxxxxxx> wrote:Dear list,I enabled quotas and set hard limits for several directories on a distribute–replicate replica 2 volume yesterday. After setting the limits I did a du, a find/stat, and a recursive ls on the paths on the FUSE mount. The usage reported in `gluster volume quota <volume> list` initially began to update, but now it's been about twelve hours and the reported usage is still much lower than actual usage.GlusterFS version 8.5 on CentOS 7. How can I force the quotas to update?Thank you,
--
--
________
Community Meeting Calendar:
Schedule -
Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
Bridge: https://meet.google.com/cpu-eiue-hvk
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
https://lists.gluster.org/mailman/listinfo/gluster-users
--
--
________ Community Meeting Calendar: Schedule - Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC Bridge: https://meet.google.com/cpu-eiue-hvk Gluster-users mailing list Gluster-users@xxxxxxxxxxx https://lists.gluster.org/mailman/listinfo/gluster-users
