What happened:
During the initial rollout of CAPMVM we noticed that the manager was crashlooping. We had only one resource to reconcile and no pre-existing clusters or any interfering objects configured in the same namespace. The logical choice was bumping the controller resources from cpu 100m /mem 30m until we felt like the controller was stable. once we reached 300m / 300m the controller began and continued without restart. We doubled every time so Im sure it can be tuned lower.
What did you expect to happen:
CAPMVM to stay up without issue
How to reproduce it:
provide a single resource after deploying on existing deployment resources.
Anything else you would like to add:
https://github.com/weaveworks-liquidmetal/cluster-api-provider-microvm/blob/64cb6accd075354a78ad31fee0129f56633dfc50/config/manager/manager.yaml#L39 but would like to performance
Environment:
6 node virtualised cluster. nodes are 4 core 16Gb RAM.
What happened:
During the initial rollout of CAPMVM we noticed that the manager was crashlooping. We had only one resource to reconcile and no pre-existing clusters or any interfering objects configured in the same namespace. The logical choice was bumping the controller resources from cpu 100m /mem 30m until we felt like the controller was stable. once we reached 300m / 300m the controller began and continued without restart. We doubled every time so Im sure it can be tuned lower.
What did you expect to happen:
CAPMVM to stay up without issue
How to reproduce it:
provide a single resource after deploying on existing deployment resources.
Anything else you would like to add:
https://github.com/weaveworks-liquidmetal/cluster-api-provider-microvm/blob/64cb6accd075354a78ad31fee0129f56633dfc50/config/manager/manager.yaml#L39 but would like to performance
Environment:
6 node virtualised cluster. nodes are 4 core 16Gb RAM.
capmvm version:
v0.8.0
OS (e.g. from
/etc/os-release):Ubuntu 20.04