[SERVER-29539] Visibility issues for readAfterClusterTime at level local Created: 09/Jun/17  Updated: 06/Dec/22  Resolved: 06/Jul/17

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

Type: Bug Priority: Major - P3
Reporter: Eric Milkie 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-29693 "lastAppliedOpTime" value can sometim... Closed
Assigned Teams:
Storage Execution
Operating System: ALL
Sprint: Storage 2017-07-31
Participants:

 Description   

When reading after a cluster time at level local, it is possible to have outstanding uncommitted transactions hide data assigned to an earlier cluster time. This is because optimes are not assigned by commit order.
To fix this, we will need to block readAfterClusterTime reads until all transactions prior to the specified cluster time have committed.
Note that it will be still be possible to return data before transactions after the specified cluster time have committed; this will be okay.


Generated at Thu Feb 08 04:21:10 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.