[SERVER-59621] Consider refactoring SBE PlanStage yield preparation & recovery logic Created: 26/Aug/21 Updated: 06/Dec/22 |
|
| Status: | Backlog |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Ian Boros | Assignee: | Backlog - Storage Execution Team |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Assigned Teams: |
Storage Execution
|
||||||||
| Participants: | |||||||||
| Description |
|
Currently, all SBE PlanStages inherit from the "CanChangeState" CRTP class, providing saveState() and restoreState() methods. Once the logic for making copies of unowned data and disabling slot access has been removed, only nodes which access storage resources will need non-trivial implementations of saveState() and restoreState(), in order to check if the collection has been dropped during a yield, for example. This gives an opportunity to refactor the saveState() and restoreState() mechanism. For example, we may want to only provide these methods on nodes which access storage resources, and remove the recursive walk of the execution tree. |
| Comments |
| Comment by Ian Boros [ 26/Aug/21 ] |
|
I've given this ticket the often controversial "Consider ..." title since this isn't strictly necessary and at this point it isn't clear what the best approach is. I'd like to leave this as a placeholder so I can spend some time trying this out. |