Hi Folks,
My lustre and kernel version: 2.6.9-55.0.9.EL_lustre.1.6.3smp
I am experiencing few problems with lustre quotas.
1)
When I run: "lfs quotacheck -ug /home" I can see in syslog lots of
LustreError messages
on mds
Nov 15 13:30:35 mds01.beowulf.cluster kernel: LustreError: 22395:0:
(ldlm_lib.c:1629:target_handle_dqacq_callback()) dqacq failed! (rc:-5)
on oss
Nov 15 13:30:35 storage07.beowulf.cluster kernel: LustreError:
15905:0:(quota_context.c:482:dqacq_completion()) acquire qunit got
error! (rc:-5)
Is there someone who can explain reason for these errors?
2)
I set only hard limit for each user. The limit is 1TB
When I run lfs quota -u <username> /home I can see that some users
have exceeded their per OST limit even if they total usage if being
far from limit. Also per OST limit seems to be not displayed
correctly. If I add up per OST limits it will not equal total limit
(see below for example)
Disk quotas for user se282 (uid 694):
Filesystem kbytes quota limit grace files quota
limit grace
/home 36479512 0 1000000000 886
0 0
ddn_home-MDT0000_UUID
120 0 102400 886 0 0
ddn_home-OST0000_UUID
13217456* 0 11366400
ddn_home-OST0001_UUID
9115004* 0 7884800
ddn_home-OST0002_UUID
8324440 0 8396800
ddn_home-OST0003_UUID
5822492 0 5939200
3) Is the MAX quota hard limit 2TB ?
I set hard limit per user to 15TB.
lfs setquota -u se282 0 15000000000 0 0 /home
Then I check user quota:
lfs quota -u se282 /home
Disk quotas for user se282 (uid 694):
Filesystem kbytes quota limit grace files quota
limit grace
/home 36479512 0 2115098112 886
0 0
ddn_home-MDT0000_UUID
120 0 102400 886 0 0
ddn_home-OST0000_UUID
13217456 0 13312000
ddn_home-OST0001_UUID
9114996 0 9216000
ddn_home-OST0002_UUID
8324440 0 8396800
ddn_home-OST0003_UUID
5822500 0 5939200
As you can see hard limit is 2115098112 instead of 15000000000
Thanks,
Wojciech Turek
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
