fix(flink): poll application deployment deletion #1611
Conversation
byashimov
force-pushed
the
byashimov-poll-flink-app-deployment-status
branch
from
ac6937d to
ef33a0a
Compare
There was a problem hiding this comment.
LGTM, but please update the test in this PR with the correct SQL, as seen in https://github.com/aiven/terraform-provider-aiven/pull/1606/files#diff-71cae1e8db8b50bfe944747bd7d61f9128e4a137878d4b00e613497bdc415f51. Will do myself in another PR.
Also please keep in mind that currently, the test succeeds with the incorrect SQL in your PR, i.e. Terraform reports that the job was created successfully, even when it ends up in an infinite restarting state, which I believe should be signalized about. Because we report that the job was created successfully to the end user, when in reality it's dead, and I think it's not great, but let's keep this discussion for another time, I think we need more opinion on this. I'm happy to merge it once the SQL is adjusted.
internal/sdkprovider/service/flink/flink_application_deployment.go
Outdated
Show resolved
Hide resolved
byashimov
force-pushed
the
byashimov-poll-flink-app-deployment-status
branch
from
ef33a0a to
a6fb011
Compare
Serpentiel
force-pushed
the
byashimov-poll-flink-app-deployment-status
branch
from
a6fb011 to
ca48019
Compare
Tests run fast and fail to delete Flink Application Deployment in transitional statuses.
The state machine for this type of resource is quite complicated. Instead of trying to replicate the transition flow this PR proposes different approach: it calls deletion over and over until succeeds or exceeds context timeout.