[SERVER-63976] TenantMigrationOplogApplier expects pre/post image noop oplog entries to have a statement id Created: 24/Feb/22  Updated: 03/Mar/23  Resolved: 03/Mar/23

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

Type: Improvement Priority: Minor - P4
Reporter: Cheahuychou Mao Assignee: [DO NOT USE] Backlog - Server Serverless (Inactive)
Resolution: Won't Do Votes: 0
Labels: shard-merge-milestone-3
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-65312 findAndModify pre/post image noop opl... Closed
related to SERVER-65344 findAndModify pre/post image noop opl... Closed
Assigned Teams:
Serverless
Sprint: Server Serverless 2022-03-21
Participants:

 Description   

The TenantMigrationOplogApplier currently expects noop pre/post image oplog entries to have a statement id. There is some inconsistency around what statement id is assigned to each pre/post image oplog entry:

  1. For the case where the side image collection is enabled, the DocumentSoureceFindAndModifyImageLookup stage sets the statement id of each forged image oplog entry to 0.
  2. For the case where the side image collection is not enable, the OpObserverImpl indirectly sets the statement id of each image oplog entry to the statement id of the oplog entry that has the pre/post image.

Pre/post image oplog entries are not linked the retryable write oplog chain, that is why re-using statement ids is not causing any issue. However, it is still confusing to make multiple oplog entries have the same statement ids, especially if the they do not belong to the same write (case 1). 



 Comments   
Comment by Esha Maharishi (Inactive) [ 03/Mar/22 ]

cheahuychou.mao, thanks for the explanation (and sorry about the delay in responding).

I don't have a good understanding of how tenant migrations handle retryable findAndModify today, but suganthi.mani is looking into this for PM-2353 next week. We'll triage this ticket in the next week or two.

Comment by Esha Maharishi (Inactive) [ 28/Feb/22 ]

elizabeth.roytburd, thanks for the heads up. Moving to the serverless backlog.

cheahuychou.mao , do you mind clarifying if this is a beginning-of-time issue with tenant migrations or something new as a result of the internal transactions API?

 

Generated at Thu Feb 08 05:59:10 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.