Open kevinlmadison opened 9 months ago
Hello @kevinlmadison , quickly looking into the logs did not spark any insights right away, hence it needs a deeper look.
I asked our MongoDB experts, they noticed that you use inMemory engine. They are asking if
writeConcernMajorityJournalDefault
was set to false
or not.
@spron-in I did some testing around the writeConcernMajorityJournalDefault
setting but this did not affect the error logs. To clarify I checked the value that we have set by default and it is true
which I discovered is not compatible with the inMemory engine. I tried 1) Setting this value to false
via rs.reconf()
and 2) Running the replicaset with the wiredTiger engine. Neither of these changes affected the error logs. Hopefully this was a good test and hopefully I described it clearly.
Thank you again!
Hello @kevinlmadison ,
so I did some digging. First, this issue has nothing to do with the engine. WiredTiger produces the same errors. Sorry for leading you in the woods.
Second, seems user assert errors started to appear since MongoDB 5. There are a couple of issues opened about it: https://www.mongodb.com/community/forums/t/user-assertions-increasing-constantly-after-upgrade-from-4-4-4-to-5-0-3/127991 https://www.mongodb.com/community/forums/t/user-asserts-on-replica-set-running-on-kubernetes/162698
In short, what MongoDB Inc suggests that it is normal behavior and that you should do is reduce the log level. As long as you don't see any issues with the cluster health - what are your thoughts about lowering the log level?
systemLog:
verbosity: 0
Hey @spron-in
Thank you so much for looking into this! With the information you've provided I think we will be able to confidently move forward. I'll use this issue to document further.
Thank you again!
Report
Hi all, I’ve been trying to debug this for a good while now, I had a pretty large mongodb deployment and I thought I had a lot of logs simply due to its size. After some digging tho it seemed that they were mostly errors. The strange thing is that the cluster still appears to work. I’ve removed most of the components including sharding, backups, and pmm but I’m still seeing the errors even with a single replicaset. I’ve also disabled istio and effectively turned off the firewall. I’m deploying this cluster using the ansible helm module, and below I’ll paste the full config and a big sample of the logs.
More about the problem
and here is a sample of the logs:
Custom resource is reporting a ready status, here are its logs:
Steps to reproduce
Versions
Anything else?
No response