[LU-3493] Need to determine if data was affected by unreadable sectors on RAID. Created: 21/Jun/13 Updated: 29/Sep/15 Resolved: 29/Sep/15 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.1.5 |
| Fix Version/s: | None |
| Type: | Task | Priority: | Minor |
| Reporter: | Herbert Yeung | Assignee: | Niu Yawei (Inactive) |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Environment: |
CentOS 6.3, SGI IS5500 (Netapp E5400), |
||
| Rank (Obsolete): | 8787 |
| Description |
|
One of our luns experienced unreadable sectors. The RAID was able to report which LBAs were affected for the lun/OST. How do we map the LBA to check if any files were affected? |
| Comments |
| Comment by Peter Jones [ 22/Jun/13 ] |
|
Niu What do you advise here? Peter |
| Comment by Andreas Dilger [ 22/Jun/13 ] |
|
If you have no partitions on your device, then the LBA (in 512-byte sectors) should be divided by 8 to get 4096-byte filesystem block numbers. Then, use "debugfs -c /dev/XXX" and the "icheck {block}[block ...]" command to map block numbers to inodes. The mapping from OST inodes to Lustre files is described in http://lists.lustre.org/pipermail/lustre-discuss/2013-April/017204.html. It may be in the manual as well (or should be added if it is not already), I didn't check. |
| Comment by Herbert Yeung [ 04/Jul/13 ] |
|
Thanks, ran through the LBA list and they were all unallocated. |
| Comment by Peter Jones [ 04/Jul/13 ] |
|
Thats good news Herbert. Is there any further action needed or shall we close out this ticket? |
| Comment by Andreas Dilger [ 04/Jul/13 ] |
|
This should be added to the user manual if it isn't already there. |