[SERVER-37207] Only retry failed writes in a batch on stale version errors in a transaction Created: 19/Sep/18  Updated: 12/Dec/23

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

Type: Task Priority: Major - P3
Reporter: Jack Mulrow Assignee: Backlog - Cluster Scalability
Resolution: Unresolved Votes: 0
Labels: LamontN, ShardedTxn:FutureOptimizations, pm-564
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
depends on SERVER-39704 Allow mongos to retry on stale versio... Needs Scheduling
Related
is related to SERVER-35707 Figure out the transaction abort stat... Closed
Assigned Teams:
Cluster Scalability
Participants:

 Description   

As of SERVER-35707, if any batch write encounters a stale shard version error, the router can only continue the transaction if this was the first overall statement received from the client. This is to avoid double applying writes because currently batch writes in a transaction must be entirely retried, unlike outside of a transaction where only the individual writes that failed will be retried. This ticket is to track changing batch write behavior in a transaction to only retry the failed writes to avoid aborting more than necessary in the face of migrations.


Generated at Thu Feb 08 04:45:20 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.