[SERVER-32424] Use WiredTiger cursor caching Created: 20/Dec/17 Updated: 30/Oct/23 Resolved: 13/Apr/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | 3.6.10, 3.7.4 |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Donald Anderson | Assignee: | Donald Anderson |
| Resolution: | Fixed | Votes: | 1 |
| Labels: | SEKB | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||||||||||||||
| Backwards Compatibility: | Fully Compatible | ||||||||||||||||||||||||||||
| Backport Requested: |
v3.6
|
||||||||||||||||||||||||||||
| Sprint: | Storage 2018-02-26, Storage Non-NYC 2018-04-09, Storage Non-NYC 2018-04-23 | ||||||||||||||||||||||||||||
| Participants: | |||||||||||||||||||||||||||||
| Linked BF Score: | 0 | ||||||||||||||||||||||||||||
| Description |
|
As part of |
| Comments |
| Comment by Githook User [ 04/Jan/19 ] |
|
Author: {'username': 'lukech', 'email': 'luke.chen@mongodb.com', 'name': 'Luke Chen'}Message: (cherry picked from commit c0d20e4fb98626143d772bb70a950b277685bbbd) |
| Comment by Githook User [ 13/Apr/18 ] |
|
Author: {'email': 'dda@mongodb.com', 'name': 'Don Anderson', 'username': 'ddanderson'}Message: |
| Comment by Alexander Gorrod [ 10/Apr/18 ] |
|
donald.anderson I believe all the pieces are in place to make this change now. Could you proceed with: Changing the default behavior so that:
Also adding a method for MongoDB users to revert to the old behavior via a configuration option. I'm hoping there is a way to tie that into the new wiredTigerCursorCacheSize option added in |
| Comment by Alexander Gorrod [ 21/Mar/18 ] |
|
The WiredTiger cursor cache has been disabled via a configuration option in |
| Comment by Alexander Gorrod [ 13/Mar/18 ] |
|
Given the scope of this change, and how sensitive some workloads are to cursor caching, I'm inclined to keep the existing cursor cache schema as a configurable startup option. That will give a reasonable escape valve if we discover any cases where the new caching scheme causes degraded performance characteristics. Hopefully that can tie into the work done in |
| Comment by Alexander Gorrod [ 08/Mar/18 ] |
|
When implementing |
| Comment by Alexander Gorrod [ 22/Feb/18 ] |
|
donald.anderson Is this a duplicate of |