[SERVER-53569] Test behavior of running data placement commands during a resharding operation Created: 04/Jan/21  Updated: 06/Dec/22  Resolved: 01/Apr/21

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

Type: Task Priority: Major - P3
Reporter: Max Hirschhorn Assignee: [DO NOT USE] Backlog - Sharding EMEA
Resolution: Won't Fix Votes: 0
Labels: PM-234-M3, PM-234-T-lifecycle
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
is related to SERVER-52760 Implement a pseudo-coordinator for Re... Closed
Assigned Teams:
Sharding EMEA
Participants:
Story Points: 2

 Description   

Chunk migration commit was made to fail if run concurrently with a resharding operation as part of SERVER-50027. This is believed to be sufficient for resharding's correctness; resharding depends on chunk ownership not changing for the duration of the operation but it doesn't depend on the chunk map staying constant. Commands such as refineCollectionShardKey and splitChunk should technically be safe if run concurrently with resharding. We should add some targeted testing to demonstrate so or add logic to prevent these commands regardless and test that behavior instead.

It would be worth auditing the list of data placement commands as part of this ticket to be doubly sure. updateZoneKeyRange is another command which had come up during the project scope.



 Comments   
Comment by Kaloian Manassiev [ 01/Apr/21 ]

This work is not necessary due to the more generic framework to disallow two DDL operations from running concurrently done under SERVER-52760.

Generated at Thu Feb 08 05:31:17 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.