Flaky: Allocator bad tls certificate#995
Conversation
markmandel
added
kind/cleanup
|
/cc @cyriltovena |
| annotations: | ||
| {{- if .Values.agones.allocator.generateTLS }} | ||
| revision/tls-cert: {{ .Release.Revision | quote }} | ||
| {{- end }} |
There was a problem hiding this comment.
Actually can you explain how this annotation is causing restart?
Should annotations be inside the if statement?
There was a problem hiding this comment.
So this puts an annotation on each pod of the deployment - with the Helm revision number.
When you update helm with a new install, the revision number increases. Since there is now a new annotation on each underlying pod in the Deployment, kubernetes will do a rolling update to the new deployment, thus restarting everything.
annotations could also be in the if statement. Should work either way (it'll be an empty annotation block otherwise). I don't mind to be honest - if you feel like it should include the annotations block, I'll change it 👍
There was a problem hiding this comment.
Thanks for the explanation and thank you so much for the fix :)
|
Build Failed 😱 Build Id: 2d07642c-1c9f-4122-bece-95ce3f157a54 To get permission to view the Cloud Build view, join the agones-discuss Google Group. |
This should fix the bad tls e2e test failures, as it will force a reload of the allocator deployment - which will load the most up to date generated ssl certificate. This is the same implementation that is in place for the controller as well for its webhook ssl certificates.
markmandel
force-pushed
the
flaky/allocator-tls
branch
from
393e037 to
0cf6a86
Compare
|
Build Succeeded 👏 Build Id: d9572719-3ddf-488b-b713-c99bee67d8ea The following development artifacts have been built, and will exist for the next 30 days:
A preview of the website (the last 30 builds are retained): To install this version:
|
|
@googlebot rescan Are you alive my bot friend? |
|
Build Succeeded 👏 Build Id: 1161ee7a-0aad-4c44-a201-6b1352e980af The following development artifacts have been built, and will exist for the next 30 days:
A preview of the website (the last 30 builds are retained): To install this version:
|
3 participants
This should fix the bad tls e2e test failures, as it will force a reload of the allocator deployment - which will load the most up to date generated ssl certificate.
This is the same implementation that is in place for the controller as well for its webhook ssl certificates.