-
Notifications
You must be signed in to change notification settings - Fork 56
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
multiMasters enabled can cause infinite wait to init in some scenarios #228
Comments
when i generate config without cert import - cert generated is puppet.puppettest.svc.cluster.local.pem (assuming its namespace specific) - this might be specific to AKS though will need more data points for other environments however i have a fix for the test that will be slightly more generic which i will open a PR for |
@anthonysomerset Thank you! Wanna become one of the official maintainers? |
I have no objections but only if you sure! - i just needed to fix use cases specific to my scenario :) i do want to look at making the restic compoent support azure blob storage as well ;) |
…aster to prevent infinite hang
I was hitting this bug too! Thanks for fixing it :) |
Describe the Bug
I am raising this as a bug because i have only been testing one scenario here and may need some other testing/data points to validate as well as to come up with a solution.
My Scenario is AKS + External Postgres (for PuppetDB)
I am importing my pre-existing infra certs and CA from legacy on prem into AKS and running multiMasters enabled.
with the default init-configmap that gets generated the check_for_masters.sh script waits indefinitely because
/etc/puppetlabs/puppet/ssl/certs/puppet.pem
is not present. it appears at least in my scenario the file that gets generated is/etc/puppetlabs/puppet/ssl/certs/puppet.puppet.svc.cluster.local.pem
I can work around the issue by executing a bash shell in the init container and then kill the script and allow it to proceed or by editing the configmap to update the
PUPPET_SSL_CERT_PEM
variable and triggering restart on the deployment at which point everything proceeds as normalI don't know if its a unique to my setup thing or if the cert is always in the FQDN format and the script never worked or what.
i did try touching an empty puppet.pem file however while this makes the default configmap work it will cause the main puppetserver container to fail at the 99-log-config.sh step - https://github.com/voxpupuli/container-puppetserver/blob/main/puppetserver/docker-entrypoint.d/99-log-config.sh because the command is only expecting one non ca.pem file present and then openssl borks because the input is broken
Expected Behavior
check_for_masters.sh should check for correct cert file and pass without issue if it is present
Steps to Reproduce
Steps to reproduce the behavior:
I am not sure if this is simply using multiMaster or alternateServerNames or the pregenerated import or a combination of all 3 - i have a suspicion that its the first 2.
here is a snippet of my values- domain changed to protect the innocent :D
Additional Context
I haven't yet created a PR here - I would be curious to find out the different possible puppserver cert names that could get generated because i think the script may need to take into account possible variations first to cover all potential scenarios - for example maybe we should attempt to read the certname field out of the puppet config and interpret correct file name from that?
The text was updated successfully, but these errors were encountered: