[SERVER-82055] Investigate and fix slow bulkWrite response cursor population on mongod Created: 10/Oct/23 Updated: 29/Oct/23 Resolved: 17/Oct/23 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | 7.2.0-rc0 |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Vishnu Kaushik | Assignee: | Sean Zimmerman |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | milestone-4 | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||||||||||
| Assigned Teams: |
Replication
|
||||||||||||||||
| Backwards Compatibility: | Fully Compatible | ||||||||||||||||
| Sprint: | Repl 2023-10-30 | ||||||||||||||||
| Participants: | |||||||||||||||||
| Description |
|
According to the flame graph, about 4% of the bulkWrite operation's time is spent in populating the cursor response on mongod (when a bulkWrite with 1000 insert operations is run). When I eliminated this method (by returning an empty cursor response instead here) I saw the corresponding improvement. Consider ways to optimize this method. And if possible, consider omitting success responses from the response and use a counter n instead, as the existing insert codepath does. See |
| Comments |
| Comment by Githook User [ 16/Oct/23 ] |
|
Author: {'name': 'seanzimm', 'email': 'sean.zimmerman@mongodb.com', 'username': 'seanzimm'}Message: |