[LU-1418] (osc_lock.c:1099:osc_lock_enqueue_wait()) DEADLOCK POSSIBLE! - too many Created: 17/May/12 Updated: 21/Nov/12 Resolved: 23/Jul/12 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | Lustre 2.3.0, Lustre 2.1.3 |
| Type: | Bug | Priority: | Minor |
| Reporter: | Alexander Boyko | Assignee: | Oleg Drokin |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Severity: | 3 |
| Rank (Obsolete): | 4538 |
| Description |
|
The console log contains 100's of "DEADLOCK POSSIBLE!" errors, followed by debug data and a stack trace. The messages sound like something is seriously wrong, but there's no sign that a deadlock has actually occurred. To the contrary, evidence indicates that the app continues executing and completes successfully. |
| Comments |
| Comment by Alexander Boyko [ 17/May/12 ] |
| Comment by Oleg Drokin [ 29/May/12 ] |
|
I see that the code in question and the comments were added by Nikita Danilov that happens to work at Xyratex at the moment. It would be really great to get some sort of a clarification from him, and perhaps some sort of a recommendation how to overcome the potential deadlock. I understand that you have not hit the deadlock yet, but it's still theoretically possible it seems, so just dropping the message does not sound like a great option. Besides I don't think I have ever seen this message before? |
| Comment by Christopher Morrone [ 18/Jun/12 ] |
|
FYI, we are also seeing hundreds of these messages in the client logs with 2.1.1-based code. Xyratex, any comments from Nikita? |
| Comment by Alexander Boyko [ 20/Jun/12 ] |
|
Not yet. |
| Comment by Alexander Boyko [ 21/Jun/12 ] |
|
Got respons from Nikita Danilov |
| Comment by Wojciech Turek (Inactive) [ 05/Jul/12 ] |
|
Just recently I installed lustre-2.1.2 on our production cluster and I seem to be getting a lot of them in the logs. If they are harmless can they be disabled? |
| Comment by Peter Jones [ 23/Jul/12 ] |
|
Landed for 2.1.3 and 2.3 |
| Comment by Nathan Rutman [ 21/Nov/12 ] |
|
Xyratex MRP-497 |