[SERVER-37262] Write background index build tests Created: 21/Sep/18 Updated: 18/Dec/18 Resolved: 18/Dec/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Eric Milkie | Assignee: | Louis Williams |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Sprint: | Storage NYC 2018-12-17, Storage NYC 2018-12-31 |
| Participants: |
| Description |
|
Write tests that confirm correct behavior when concurrent writes happen behind and ahead of collection scan cursor. |
| Comments |
| Comment by Louis Williams [ 18/Dec/18 ] |
|
Closing in favor of tests written for |
| Comment by Daniel Gottlieb (Inactive) [ 08/Nov/18 ] |
|
This ticket is to give confidence that inserts that can legally be missed by a table scanning cursor are correctly inserted into the "side writes" table. I don't believe any (current) implementations can naturally insert being a collection scanning cursor. One hack/trick that may make this testable for low-effort is to introduce a failpoint/test only server parameter that turns the record store table scan into a walk on the _id index. The client performing writes concurrent to the index build can then choose a random _id for performing inserts. |