[SERVER-17696] Terminate sharded queries immediately after a failure Created: 23/Mar/15 Updated: 06/Dec/22 Resolved: 23/Apr/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Osmar Olivo | Assignee: | Backlog - Query Team (Inactive) |
| Resolution: | Duplicate | Votes: | 3 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||||||||||||||
| Assigned Teams: |
Query
|
||||||||||||||||||||||||||||
| Participants: | |||||||||||||||||||||||||||||
| Case: | (copied to CRM) | ||||||||||||||||||||||||||||
| Description |
|
If a sharded query incurs a failure (such as a dead cursor) on any one shard, when that error is returned to the mongos, mongos should terminate the query on the remaining shards automatically. Otherwise, we will continue utilize processing power and time on a query we already know is going to fail. This is particularly an issue if you have many long running cpu intensive queries, such as for analytics. I would like a failed query to be cleaned up across the cluster as fast as possible, or within some reasonable timeout, in order to free up resources more quickly for other queries. |
| Comments |
| Comment by David Storch [ 23/Apr/18 ] |
|
For more details about this improvement, see the issue description of |
| Comment by Charlie Swanson [ 23/Apr/18 ] |
|
This was implemented in |