Use retry logic when `kapp deploy` encounters `client connection lost`
idoru opened this issue · 1 comments
Describe the problem/challenge you have
kapp deploy
fails without retry on http2: client connection lost
Example from our CI environment which uses kapp deploy:
| 3:45:49PM: ---- waiting on 6 changes [61/67 done] ----
3:46:49PM: Deploy failed
| E0607 15:46:48.669366 32 request.go:1058] Unexpected error when reading response body: http2: client connection lost
| kapp: Error: waiting on reconcile packageinstall/bitnami-services (packaging.carvel.dev/v1alpha1) namespace: tap-install:
| Errored:
| Getting resource packageinstall/bitnami-services (packaging.carvel.dev/v1alpha1) namespace: tap-install:
| API server says: unexpected error when reading response body. Please retry. Original error: http2: client connection lost (reason: )
| Deploying: Error (see .status.usefulErrorMessage for details)
Describe the solution you'd like
Use retry logic when this sort of network error is encountered.
Anything else you would like to add:
N/A
Vote on this request
This is an invitation to the community to vote on issues, to help us prioritize our backlog. Use the "smiley face" up to the right of this comment to vote.
👍 "I would like to see this addressed as soon as possible"
👎 "There are other more important things to focus on right now"
We are also happy to receive and review Pull Requests if you want to help working on this issue.
We do have a set of errors we retry for, this would qualify as an addition if we are sure that we won't risk retrying in cases which we cannot recover from while matching to this error string