[LU-4887] LFSCK default behavior should be scan all known types Created: 12/Apr/14 Updated: 20/Jun/14 Resolved: 20/Jun/14 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.6.0, Lustre 2.5.1 |
| Fix Version/s: | Lustre 2.6.0 |
| Type: | Improvement | Priority: | Critical |
| Reporter: | Richard Henwood (Inactive) | Assignee: | Andreas Dilger |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||
| Rank (Obsolete): | 13523 | ||||||||
| Description |
|
Running LFSCK is not obvious, even for users like myself that know a lot about it. Running lctl lfsck_start -A -M fsname-MDT0000 reports that LFSCK was started, but it only starts the OI Scrub when no "-t" options are specified. I think this is confusing even for LFSCK-knowledgeable people such as myself. It took Oleg and I a few minutes to figure out why LFSCK was not reporting any progress via lctl get_param mdd.*.lfsck* and repeated lfsck_start commands, and I think i would be much harder for normal users to figure out why their filesystem is not being checked. It is possible that users won't even know that the namespace and layout checks are not being run. If no -t option is given, or -t all is given, then all of the known LFSCK checks should be run. If specific checks are listed such as -t namespace or -t layout then only those checks should be run. If -t scrub is given, then only the OI Scrub should be run as is currently the default. |
| Comments |
| Comment by Andreas Dilger [ 09/May/14 ] |
|
Nasf, I'm actually working in a patch for this. |
| Comment by Andreas Dilger [ 09/May/14 ] |
| Comment by nasf (Inactive) [ 08/Jun/14 ] |
|
The following two patches are for interoperability tests, and should be landed to related branches after the patch http://review.whamcloud.com/10278 landed. http://review.whamcloud.com/10641 (b2_5) |
| Comment by Peter Jones [ 20/Jun/14 ] |
|
Landed for 2.6 |