Huge mongodb.log on Percona MongoDB Aribiter

MongoDB Percona Server for MongoDB
1

Hi All,

I have a Percona MongoDB (version 7.0.11-6) replica set consisting of two data-bearing nodes and one arbiter node.

While reviewing the arbiter’s log (mongodb.log), I noticed several lines (250K rows in 3 days) with information like the following:
{“t”:{“$date”:“2024-10-10T03:43:56.002+02:00”},“s”:“W”, “c”:“QUERY”, “id”:23799, “ctx”:“ftdc”,“msg”:“Aggregate command executor error”,“attr”:{“error”:{“code”:26,“codeName”:“NamespaceNotFound”,“errmsg”:“Unable to retrieve storageStats in $collStats stage :: caused by :: Collection [local.oplog.rs] not found.”},“stats”:{},“cmd”:{“aggregate”:“oplog.rs”,“cursor”:{},“pipeline”:[{“$collStats”:{“storageStats”:{“waitForLock”:false,“numericOnly”:true}}}],“$db”:“local”}}}

It seems there is an error related to the execution of the Aggregate command, specifically a NamespaceNotFound error, which is likely caused by the missing local.oplog.rs collection.

Has anyone encountered a similar issue or have suggestions on how to resolve this?

Thanks in advance for your help!

AT

2

Hello,
Arbiters are not supposed to have any collections or execute any aggregations.
I would suggest to check replica set status information and ensure there is nothing wrong with this arbiter’s config. Then if configuration is OK try to figure out from the logs which client is trying to execute aggregation on arbiter.

3

I have the same problem with the exactly same configuration with 2 replicas + 1 arbiter.

4

Hi Max,
while analyzing the mongodb.log, I noticed numerous entries related to FTDC (Full-Time Diagnostic Data Capture) with the context “ctx”:“ftdc”.
To reduce these entries, I modified the MongoDB configuration file adding the following setting (only on arbiter node):

setParameter:
diagnosticDataCollectionEnabled: false

After applying this change, mongodb.log is smaller