[SERVER-59518] Change query yielding behavior to WT::commit_transaction instead of WT::rollback_transaction Created: 23/Aug/21  Updated: 06/Dec/22  Resolved: 23/Aug/21

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Major - P3
Reporter: Dianna Hohensee (Inactive) Assignee: Backlog - Storage Execution Team
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Duplicate
duplicates SERVER-59517 POC: change query yielding to retain ... Closed
Assigned Teams:
Storage Execution
Participants:

 Description   

The following is how the current code path for yielding works:

Query yield releases the storage engine resources here. Locked reads will implicitly open a new storage transaction on the next storage access; lock-free reads will explicitly open a new storage transaction in the restore logic here.
abandonSnapshot() then closes the storage transaction here.
_txnClose then calls rollback_transaction here.

One quick and dirty solution to pivot to calling WT::commit_transaction instead of WT::rollback_transaction would be a bool decoration on the opCtx set during the query yield logic. That should get things up and running, to find out if there's a performance issue. Then a better solution could be explored. But maybe there'll be an obvious good way to do it.


Generated at Thu Feb 08 05:47:27 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.