[LU-3843] osc_completion()) ASSERTION( (!!(page->cp_state == CPS_PAGEOUT) == !!(cmd == 0x02)) Created: 27/Aug/13 Updated: 19/Mar/14 Resolved: 19/Mar/14 |
|
| Status: | Closed |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | Lustre 2.6.0 |
| Type: | Bug | Priority: | Major |
| Reporter: | Christopher Morrone | Assignee: | Niu Yawei (Inactive) |
| Resolution: | Cannot Reproduce | Votes: | 0 |
| Labels: | None | ||
| Environment: |
Sequoia login node. PPC64. 48 core Power 7. Lustre version 2.4.0-RC2_11chaos |
||
| Severity: | 3 |
| Rank (Obsolete): | 9945 |
| Description |
|
Using Lustre version 2.4.0-RC2_11chaos, we hit the following assertion on a login node (lustre client): osc_cache.c:1267:osc_completion()) ASSERTION( (!!(page->cp_state == CPS_PAGEOUT) == !!(cmd == 0x02) ) failed: The backtrace looks roughly like this: osc_ap_compltion osc_extent_finish brw_interpret ptlrpc_check_set ptlrpcd_check ptlrpcd There was not much going on in the client's log. About one minute earlier an OSC lost connection to one of the OSTs and then reconnected. But that is a fairly common occurance. It is not immediately obvious to me that there is a connection to this assertion. This client is a Power7 login node with 48 cores and 64GB of RAM. |
| Comments |
| Comment by Peter Jones [ 27/Aug/13 ] |
|
Niu Could you please comment on this one? Thanks Peter |
| Comment by Niu Yawei (Inactive) [ 29/Aug/13 ] |
|
I have no idea how this can be triggered now, maybe printing more information on the LASSERT could be helpful for debug: http://review.whamcloud.com/7494 Chris, could you apply this debug patch? so we can see the value of cp_state & cmd next time when it's triggered. Thanks. |
| Comment by Christopher Morrone [ 30/Aug/13 ] |
|
I had the same thought. Yes, I will add that to our local tree. |
| Comment by Peter Jones [ 02/Nov/13 ] |
|
Landed for 2.6 |
| Comment by Niu Yawei (Inactive) [ 04/Nov/13 ] |
|
The landed patch is just a debug patch. |
| Comment by Jodi Levi (Inactive) [ 19/Mar/14 ] |
|
Debug patch has landed. If this occurs again, please reopen the bug. |