[SERVER-55136] Investigate BucketCatalog behavior when write to underlying document fails Created: 10/Mar/21  Updated: 29/Oct/23  Resolved: 25/Mar/21

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: None
Fix Version/s: 5.0.0-rc0

Type: Bug Priority: Major - P3
Reporter: Dan Larkin-York Assignee: Gregory Noma
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Backwards Compatibility: Fully Compatible
Sprint: Execution Team 2021-04-05
Participants:

 Description   

We may currently have undefined behavior in the case that the write to the underlying bucket document fails. Currently we handle the update-a-missing-document case by simply clearing the bucket from the BucketCatalog and aborting any pending writes to that bucket. But any other error case seems to be ignored, and the bucket is left with some (likely invalid) state and it's allowed to be used for future measurement insertions. We may need to either clear the bucket in this case as well or think rather carefully about the bucket state (stats, and especially min/max) and how to manage it. At the very least, we need to test such cases, and possibly fix them.



 Comments   
Comment by Githook User [ 25/Mar/21 ]

Author:

{'name': 'Gregory Noma', 'email': 'gregory.noma@gmail.com', 'username': 'gregorynoma'}

Message: SERVER-55136 Clean up invalid state on failed time-series insert
Branch: master
https://github.com/mongodb/mongo/commit/75d3a439ea77e6f8bd9bed2b4f397803b66c2949

Generated at Thu Feb 08 05:35:34 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.