[LU-8620] Get rid of mystery object '???<object#138>' from the output of zdb Created: 16/Sep/16 Updated: 08/Feb/18 Resolved: 08/Feb/18 |
|
| Status: | Closed |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Minor |
| Reporter: | Jinshan Xiong (Inactive) | Assignee: | Jinshan Xiong (Inactive) |
| Resolution: | Not a Bug | Votes: | 0 |
| Labels: | None | ||
| Issue Links: |
|
||||||||
| Rank (Obsolete): | 9223372036854775807 | ||||||||
| Description |
|
It turned out the objects are object accounting objects and somehow they are created under MASTER_NODE_OBJ. We should create them under root directory of dataset. |
| Comments |
| Comment by Andreas Dilger [ 16/Sep/16 ] |
|
This specific object will become obsolete as soon as the ZFS dnode accounting feature is available for use. However, I think that there may be other ZFS OSD objects that similarly do not have a proper pathname (only a FID), but I don't have any specific example right now. |
| Comment by Gerrit Updater [ 16/Sep/16 ] |
|
Jinshan Xiong (jinshan.xiong@intel.com) uploaded a new patch: http://review.whamcloud.com/22543 |
| Comment by Jinshan Xiong (Inactive) [ 16/Sep/16 ] |
|
Yes, all dynamically created object with seq number ```FID_SEQ_LOCAL_NAME``` don't have names. |
| Comment by Andreas Dilger [ 16/Sep/17 ] |
|
I think (most of) the nameless objects in osd-zfs are being fixed as part of ZFS OI Scrub patch https://review.whamcloud.com/28610 |
| Comment by Joseph Gmitter (Inactive) [ 07/Feb/18 ] |
|
Is this work still needed? |