[SERVER-2782] no rollback of chunking if chunking fails (e.g. for large collections) Created: 17/Mar/11 Updated: 03/Jun/19 Resolved: 03/Jun/19 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Greg Studer | Assignee: | Kaloian Manassiev |
| Resolution: | Duplicate | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Operating System: | ALL | ||||||||
| Participants: | |||||||||
| Description |
|
When chunking fails (e.g. for very large collections) the chunk data remains in the config database, causing hard-to-debug errors when sharding is tried again. Fix is to rollback sharding when chunking fails. May also be an issue if unsharding is ever implemented. FIx to review @ http://crucible.10gen.com/cru/SERVER-221 |
| Comments |
| Comment by Eliot Horowitz (Inactive) [ 18/Sep/12 ] |
|
Note: the way initial sharding works has changed a lot, so this is likely a pretty minor issue now. |
| Comment by Greg Studer [ 17/Mar/11 ] |
|
Same issue occurs when sharding fails because of non-unique indices, fairly annoying. |