[SERVER-37742] Avoid unnecessary shard version metadata refreshes Created: 24/Oct/18 Updated: 29/Oct/23 Resolved: 05/Nov/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | None |
| Fix Version/s: | 4.1.5 |
| Type: | Bug | Priority: | Major - P3 |
| Reporter: | Matthew Saltz (Inactive) | Assignee: | Natalie Tsvetkova |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | sharding-wfbf-day | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Backwards Compatibility: | Fully Compatible |
| Operating System: | ALL |
| Sprint: | Sharding 2018-11-05, Sharding 2018-11-19 |
| Participants: |
| Description |
|
We're missing a `return Status::OK()` here that leads us to do refreshes even when we could be short-circuiting early. |
| Comments |
| Comment by Githook User [ 05/Nov/18 ] |
|
Author: {'name': 'Natalie Tsvetkova', 'email': 'natalie.tsvetkova@mongodb.com', 'username': 'ntsvetko'}Message: |
| Comment by Kaloian Manassiev [ 31/Oct/18 ] |
|
ramon.fernandez, no this is a much more benign check which short-cuts refreshing if the router has a staler version knowledge than the shard and is already like that on 4.0. |
| Comment by Kaloian Manassiev [ 25/Oct/18 ] |
|
Good catch - looks like it came from this commit. Since it's a one line change to add a return - can you do it? It will take less time than us having to triage it in the meeting. See also if it needs to be fixed in 4.0. |