MongoDB 7.0.5 Community on Openshift 4.13.44

Ops and Admin
,
1

Dear Community,

We have recently setup MongoDB 7.0.5 Community on Openshift 4.13.44 using the bitnami helm-charts.

We basically have an HA Cluster replica of 3, with an additional HA Hidden Read Only Replica that acts as a slave.

We are noticing that the Hidden Replicas are constantly claiming memory that never gets returned back, especially when running operations such as mongodump.

I’ve already tried setting --wiredTigerCacheSizeGB=1

Any ideas to what might be wrong?

2

I’d like to add a bit more information to this:

  • The MongoDB that is being backed up is approximately 3.4GB when Compressed
  • The Container has 2GB Requested Memory and 4GB Limit
  • I’ve set the wiredTigerCacheSizeGB=2

This is the argument I’m passing to my mongodump:
mongodump --host $MONGO_URI --username=$MONGO_ROOT_USERNAME --password=$MONGO_ROOT_PASSWORD --authenticationDatabase=admin
–gzip --readPreference=secondary --numParallelCollections=1

Here’s what I’m seeing:

  • Hidden Replica is restarted
  • First backup is taken, and during the entire backup the 2GB Memory is not exceeded
  • Second backup is taken, memory starts climbing

MongoBackup
MongoBackup698×330 13.6 KB