Description:
I currently have few mongo databases on EKS with Everest 1.9.0. While trying to create a scheduled backup/adhoc backup using S3( the aws bucket is configured) we can observe the backup is triggered, the mongo pod logs(backup container) shows that command received and starts the backup however it fails in a few minutes with an OOMKilled message on the pod for the backup container. The backup job then fails and backup is not created.
Steps to Reproduce:
The backup container has a measly 500m cpu and 1G memory which seems to insufficient for our 50-60GB database. I’m unable to provision more resources for this container specifically. I’ve tried to update the PerconaServerMongoDB CR or the stateful set that this mongodb created - they both get overwritten. The DatabaseCluster CR doesn’t seem to support resource allocation for the container. If you could suggest where i can make this change to increase this backup container resources for future mongo databases created over Everest UI it would be helpful. This is currently blocking some important tests. Thanks
Version:
v1.9.0 ( also had the same issue on v1.7.0 and v1.8.0 and v1.8.1. Upgrading has not helped.