Hello, we are having a problem that I imagine is due to some basic configuration that I might be missing. Currently, I am using version 6 of MongoDB and migrating to version 8. I am experiencing a behavior where, in version 8, MongoDB utilizes all the machine’s memory, and when it reaches the limit, the MongoDB service stops. In version 6, the same query and the same database do not cause this issue.
I have already checked the ‘allowDiskUse: true’ variable in the query and also the ‘storage: wiredTiger: engineConfig: cacheSizeGB: 12’ variable properly configured, but without success. The service still crashes. Do you have any tips for adjustments?
1 Like
One more thing. Looks like it only happens in shard clusters. We don’t observe this behaviour in our replica sets.
After upgrading to 8 we also experience the same problem. In our case it happens only to primary members (we have 10 shards.) Over time the instance memory keeps getting hogged by the MongoDB process, and finally the instance just stops working because all the memory is used.