[SERVER-15782] repairDatabase should not require twice the disk space in use by databases Created: 22/Oct/14 Updated: 10/Nov/15 Resolved: 22/May/15 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage |
| Affects Version/s: | 2.6.4 |
| Fix Version/s: | None |
| Type: | New Feature | Priority: | Major - P3 |
| Reporter: | Laurent Glayal | Assignee: | Unassigned |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Participants: |
| Description |
|
Hi, when a lot of documents are inserted then deleted from a database the compaction can be done but disk space in use is not freed back to OS. The repairDatabase always require a lot of free space on disk to work, even if the size of remaining documents and indexes is a fraction of initial files sizes. repairDatabase or equivalent should be able to require free disk space to proceed using the size of remaining data and not the one of current data files. |
| Comments |
| Comment by Laurent Glayal [ 01/Jun/15 ] |
|
Does it mean newer storage will workaround this issue ? |
| Comment by Geert Bosch [ 22/May/15 ] |
|
The MMAPv1 engine intentionally uses an out-of-place repair mechanism to ensure atomicity of the repair action. |