[SERVER-81582] Transaction limit can be hit in addShard Created: 29/Sep/23  Updated: 07/Feb/24

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

Type: Bug Priority: Major - P3
Reporter: Marcos José Grillo Ramirez Assignee: Backlog - Catalog and Routing
Resolution: Unresolved Votes: 0
Labels: oldshardingemea
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Assigned Teams:
Catalog and Routing
Operating System: ALL
Participants:

 Description   

SERVER-71755 added code to insert the shard metadata along with the database metadata and the placement history in a transaction as part of PM-1939, this fixes addShard not registering all the databases if there is a failure in the middle of the write operation, and allows a stable timestamp to get the necessary metadata for change streams, however, a replica set with too many databases could have a similar problem as the one described in SERVER-58433, where we reach the limits of the WiredTiger cache size for a transaction.

SERVER-79160 adds to that transaction the collections present in the shard, making this problem more probable. We should batch the request in several transactions to mitigate this, until we have a fully reliable addShard which should come in PM-2050.


Generated at Thu Feb 08 06:46:53 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.