[SERVER-50048] minValid on another branch of history is unrecoverable but not detected Created: 31/Jul/20  Updated: 17/Aug/20  Resolved: 17/Aug/20

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

Type: Bug Priority: Major - P3
Reporter: Matthew Russotto Assignee: Matthew Russotto
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
Duplicate
duplicates SERVER-49430 Running hang analyzer in awaitSeconda... Closed
Related
is related to SERVER-43867 Work around unrecoverability of rollb... Closed
Operating System: ALL
Sprint: Repl 2020-08-10, Repl 2020-08-24
Participants:
Linked BF Score: 20

 Description   

With enable-majority-read-concern false (EMRC=false), we may have a failure during rollback, when the rollback source was on a branch of history which is ultimately rolled back itself. This results in the minValid document on the failed node being on the wrong branch of history. If no sync source still has this branch of history, the situation is unrecoverable, but we do not detect this.

It's not 100% clear how we would; we could see if no nodes had the requiredOpTime and if so declare it unrecoverable, but this wouldn't solve the BF because one node was unreachable at the time.


Generated at Thu Feb 08 05:21:34 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.