[LU-11856] Spurious "accounting not enabled: enable feature@userobj_accounting in pool" message Created: 14/Jan/19 Updated: 25/Nov/19 Resolved: 12/Jul/19 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Jeff Johnson (Inactive) | Assignee: | Nathaniel Clark |
| Resolution: | Cannot Reproduce | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||
| Severity: | 3 | ||||
| Rank (Obsolete): | 9223372036854775807 | ||||
| Description |
|
Upon upgrading spl/zfs to 0.7.9 the existing pools needed to be upgraded and were upgraded to the new 0.7x on-disk format. When mounting the MDT the following error message occurs: Lustre: 44469:0:(osd_handler.c:1196:osd_mount()) ls15-mds-00.mdt/mdt: dnode accounting not enabled: enable feature@userobj_accounting in pool The state of the MDT pool is: ls15-mds-00.mdt multihost on local ls15-mds-00.mdt feature@async_destroy enabled local ls15-mds-00.mdt feature@empty_bpobj active local ls15-mds-00.mdt feature@lz4_compress active local ls15-mds-00.mdt feature@multi_vdev_crash_dump enabled local ls15-mds-00.mdt feature@spacemap_histogram active local ls15-mds-00.mdt feature@enabled_txg active local ls15-mds-00.mdt feature@hole_birth active local ls15-mds-00.mdt feature@extensible_dataset active local ls15-mds-00.mdt feature@embedded_data active local ls15-mds-00.mdt feature@bookmarks enabled local ls15-mds-00.mdt feature@filesystem_limits enabled local ls15-mds-00.mdt feature@large_blocks enabled local ls15-mds-00.mdt feature@large_dnode enabled local ls15-mds-00.mdt feature@sha512 enabled local ls15-mds-00.mdt feature@skein enabled local ls15-mds-00.mdt feature@edonr enabled local ls15-mds-00.mdt feature@userobj_accounting active local |
| Comments |
| Comment by Peter Jones [ 14/Jan/19 ] |
|
Nathaniel Can you please investigate? Thanks Peter |
| Comment by Nathaniel Clark [ 21/Jan/19 ] |
|
I cannot generate this message spuriously. Formatting MGT/MDT with lustre 2.9.0 ZFS 0.6.5.9 then importing to Lustre 2.10.6 ZFS 0.7.9. I can generate the message correctly when first mounting, but after running zpool upgrade POOL, the device mounts without the error message. I even tried with -o nosvc to see if that had some affect, to no avail. Do you have a set of reproduction steps? |