[SERVER-24826] Pending chunk can be removed from the recipient shard while applyOps could still succeed Created: 28/Jun/16  Updated: 06/Dec/22  Resolved: 17/Dec/21

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

Type: Bug Priority: Major - P3
Reporter: Dianna Hohensee (Inactive) Assignee: [DO NOT USE] Backlog - Sharding Team
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
Assigned Teams:
Sharding
Operating System: ALL
Participants:

 Description   

The scenario under which this bug can occur is:

  • Donor starts migration of chunk C to Recipient at which point the recipient will have C on the pending list.
  • Donor sends applyOps command to the config server, which blocks for some reason before it reaches the the CSRS primary.
  • With the applyOps command not having run yet, the recipient is restarted, at which point it will forget about the pending chunk. At this point the recipient can potentially run cleanupOrphans, which would wipe out the contents of C.
  • Donor's applyOps command finally runs and the hand-off of chunk to recipient is successful at which point the contents of that chunk would be lost.


 Comments   
Comment by Kaloian Manassiev [ 17/Dec/21 ]

This is gone away as result of PM-812.

Generated at Thu Feb 08 04:07:31 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.