[SERVER-29105] Make _mergeAuthzCollections ignore MaxTimeMS Created: 08/May/17  Updated: 06/Dec/22

Status: Open
Project: Core Server
Component/s: Internal Code
Affects Version/s: None
Fix Version/s: features we're not sure of

Type: Task Priority: Major - P3
Reporter: Spencer Jackson Assignee: Backlog - Security Team
Resolution: Unresolved Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Assigned Teams:
Server Security
Participants:

 Description   

_mergeAuthzCollections is an internal command, which performs a very specific, and expensive, operation for mongorestore.

mongos may attach a MaxTimeMS to this command, when it proxies it to the config servers. This can cause restoring many user documents to fail.

Because _mergeAuthzCollections is internal, requires extensive privileges, performs an intentionally expensive operation, and failing due to timeout seems undesirable, it should ignore MaxTimeMS.



 Comments   
Comment by Spencer Brody (Inactive) [ 12/May/17 ]

I think the better solution would be for mongos to not always send maxTimeMs for _mergeAuthzCollections, but I'm not sure how easy that is to do with the current sharding code. Last I was aware we were adding maxTimeMs to every command sent to the config servers no matter what at a pretty low level that would be hard to circumvent, but I know the sharding team's been touching that code a bunch recently so perhaps it's gotten easier now.

Comment by Andy Schwerin [ 09/May/17 ]

I'd like to discuss this. I'm not thrilled by commands deciding they're too important to ignore maxTimeMS, and would like to understand the scenario better.

Generated at Thu Feb 08 04:19:55 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.