RSC: Resetting secret key?

[pat-s]

Assuming one has an asynced state of the secret key between the DB and the RSC instance, how would one resolve it for k8s deployments?

I am aware of https://docs.rstudio.com/connect/admin/appendix/cli/#rscadmin and rscadmin --reset-secret-key. However, in this state the RSC pod does not even start and hence one cannot run rscadmin to solve the issue (as on a host or docker install where the startup can more easily be overriden).

Is there an easy solution which I am overlooking?

[colearendt]

Oops, apologies for the delay getting back to you here!! One simple (ugly) solution is to set sleep as your command:

command: ["sleep"]
args: ["infinity"]
readinessProbe:
  enabled: false

However, we are tracing through some work that we did on a project called rskey that makes it easier to generate this value elsewhere and then make use of it in helm values / etc.

https://github.com/rstudio/rskey

Getting this working definitely involves a bit of nuance that we have not yet documented adequately IMO. If you're interested in using rskey, I would definitely be happy to help get you the order of operations though!

[pat-s]

Thanks for the workaround! Looking forward to the integration of rskey!

[dbkegley]

Closing this since the solution falls outside of the helm chart. The best practice here is to use a kubernetes Secret to manage this credentials independently of the Connect process with rskey. If they need to be encrypted then you should enable a secret manager in the cluster