[SERVER-23418] remove_and_bulk_insert.js failing on PowerPC Created: 30/Mar/16 Updated: 18/Nov/16 Resolved: 01/Apr/16 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | WiredTiger |
| Affects Version/s: | None |
| Fix Version/s: | 3.2.5, 3.3.4 |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Michael Cahill (Inactive) | Assignee: | Geert Bosch |
| Resolution: | Done | Votes: | 1 |
| Labels: | code-only | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Backwards Compatibility: | Fully Compatible | ||||||||
| Operating System: | ALL | ||||||||
| Backport Completed: | |||||||||
| Steps To Reproduce: | Run fsm_all_replication.js in one of the concurrency suites on PowerPC:
|
||||||||
| Sprint: | Integration 13 (04/22/16) | ||||||||
| Participants: | |||||||||
| Linked BF Score: | 0 | ||||||||
| Description |
|
The test concurrency/fsm_workloads/remove_and_bulk_insert.js is failing on PowerPC. That test has 5 threads, each doing a mix of bulk inserts and removing all documents in the same collection. The failure is that one of the removes tries to delete a document that is not visible, leading to a WT_NOTFOUND. If there was a yield or write conflict between getNext returning a RecordID and the deleteDocument call for that ID, a WT_NOTFOUND from WiredTiger is possible because a concurrent remove could have completed in the meantime. We see lots of write conflicts and yields in this workload. Attempts to reproduce on x86 have so far been unsuccessful. |
| Comments |
| Comment by Githook User [ 01/Apr/16 ] |
|
Author: {u'username': u'GeertBosch', u'name': u'Geert Bosch', u'email': u'geert@mongodb.com'}Message: (cherry picked from commit 2ab9b1fd7ea20319e2d9ddb0532234729877703f) |
| Comment by Githook User [ 01/Apr/16 ] |
|
Author: {u'username': u'GeertBosch', u'name': u'Geert Bosch', u'email': u'geert@mongodb.com'}Message: |
| Comment by Ernie Hershey [ 31/Mar/16 ] |
|
michael.cahill - I can take one of the Power RHEL VM's out of Evergreen temporarily and give you shell access. Mark B. used one to analyze the core dump and it worked well. Can you let me know when? I opened BUILD-1401 for record keeping. |