[LU-1782] Ignore sb_has_quota_active() in OFED's header

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.1.4, Lustre 1.8.9
• Affects Version/s: Lustre 2.1.4, Lustre 1.8.8
• Environment: b1_8, out-of-kernel-tree OFED, RHEL5
• Severity: 3
• Assignee: Niu Yawei (Inactive)
• Reporter: Shuichi Ihara (Inactive)

Description

On b1_8 + out-of-kerneltree OFED (e.g. Mellanox 1.5.3), the quota doesn't work well.

1 x MDT, 1 x OST are mounted on a server and one client mounts that Lustre.

1. lfs quotaon -ug /l-exofed/
2. lfs quotacheck /l-exofed
3. lfs setquota -B 0 -g group1 /l-exofed/
4. lfs setquota -B 100000 -g group1 /l-exofed/
5. lfs quota -g group1 -v /l-exofed/
   Disk quotas for group group1 (gid 1001):
   Filesystem kbytes quota limit grace files quota limit grace
   /l-exofed/ 0 0 100000 - 0 0 0 -
   lustre-MDT0000_UUID
   0 - 1 - 0 - 0 -
   lustre-OST0000_UUID
   0 - 1 - - - - -

1. su - user1
   -bash-3.2$ cd /l-exofed/
   -bash-3.2$ mkdir a
   mkdir: cannot create directory `a': Disk quota exceeded

The problem seems to be caused by two functions sb_has_quota_active() and sb_any_quota_active() defined in
$BACKPORT_INCLUDES/include/linux/quotaops.h and use them, instead of lustre defined.
1.8.8 was no problem with out-of-kerneltree OFED, so LU-1438 patches might be related.

I created the patches and will post it soon.

Issue Links

• is related to: LU-340 system hang when running sanity-quota on RHEL5-x86_64-OFED
• Trackbacks: Lustre 1.8.x known issues tracker While testing against Lustre b18 branch, we would hit known bugs which were already reported in Lustre Bugzilla https://bugzilla.lustre.org/. In order to move away from relying on Bugzilla, we would create a JIRA