[SERVER-29388] Investigate changing default maximum in-memory page size with WiredTiger Created: 26/May/17 Updated: 06/Dec/22 Resolved: 24/Jun/19 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Storage, WiredTiger |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Alexander Gorrod | Assignee: | Backlog - Storage Execution Team |
| Resolution: | Incomplete | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||
| Assigned Teams: |
Storage Execution
|
||||
| Participants: | |||||
| Description |
|
MongoDB currently configures the WiredTiger memory_page_max to be 10MB for all collections. That setting tells WiredTiger how much memory to accumulate in a page before splitting or reconciling it. There is a performance trade off for choosing different sizes, but 10MB does not work very well when MongoDB users insert documents that are larger than 5MB. We should create several different workloads that are likely to benefit from different settings, and use those workloads to decide if the current default is the best value. The tests should be run on both fast and slow I/O subsystems. |
| Comments |
| Comment by Sara Williamson [ 24/Jun/19 ] |
|
We don't believe this is a problem at this time. |
| Comment by Alexander Gorrod [ 26/May/17 ] |
|
Note that there is some discussion of the best setting to be using in |