[SERVER-57482] Adaptively call wiredtiger_calc_modify Created: 06/Jun/21 Updated: 09/Aug/23 |
|
| Status: | Backlog |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Improvement | Priority: | Major - P3 |
| Reporter: | Yuhong Zhang | Assignee: | Backlog - Storage Execution Team |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | time-series | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||
| Assigned Teams: |
Storage Execution
|
||||||||||||||||
| Backport Requested: |
v5.0
|
||||||||||||||||
| Participants: | |||||||||||||||||
| Description |
|
When updating in wiredtiger record store, wiredtiger_calc_modify is called to attempt to get a diff for faster update. However, when it fails to calculate such a diff, the time goes wasted. For certain workload (e.g. time series), we found that the calculation rarely succeeds and slows down the updates. We would like to implement a general heuristic (not just for time series) to adaptively increase or decrease the frequency on attempting to do the calculation. After implementing the heuristic, the "engineering solution" for time-series in the linked ticket should be removed. |
| Comments |
| Comment by Henrik Edin [ 23/Aug/21 ] |
|
Consider adding statistics for how often this succeeds |