fix: switching to the registry addon for olm testing (#40334) #40349
Conversation
|
Seems there is a related error: |
My only conclusion is that every single part of Docker is awful... I've rerun the job to help clarify if this is error is deterministic. The only reference for something like this is in https://minikube.sigs.k8s.io/docs/handbook/vpn_and_proxy/ which mentions a proxy blocking access. I can also:
|
closes: keycloak#40099 Signed-off-by: Steve Hawkins <shawkins@redhat.com> (cherry picked from commit eb96b4a)
51df5bd to
602a7d5
Compare
|
What if we used podman as the minikube driver instead? :D |
It's still considered experimental of course - and we'd still need to run a registry... Since the second run passed we can rule out a proxy issue. I also see no reason why the ip would change given what I've seen locally and with every other test run, but just in case I've changed it to use a subnet and minikube ip - along with a print out of the ip - that thrid time ran as well, and confirmed the usage of 192.168.49.2. I've set it run repeatedly to see how often we might encounter this problem. |
d2d0238 to
602a7d5
Compare
|
@vmuzikar the olm tests run 53 more times without a problem, so I'm thinking we are safe to merge even if we've only reduced, but not fully eliminated the flakiness. |
|
@shawkins Agree, let's merge it. Could you please prepare a PR to align the changes in |
I can, but there was no indication from the additional runs that a different ip is being used - they were done still with a fixed ip as the registry. It also won't prevent an issue if we switch to a different driver. We'd have to instead look at all the possible default ip assignments ranges and proactively add those to the insecure-registry settings. |
|
@shawkins Yeah, seems like the IP is actually stable. But I'd still prefer to have the branches aligned with the same workflow for the minikube. |
…keycloak#40349) closes: keycloak#40099 (cherry picked from commit eb96b4a) Signed-off-by: Steve Hawkins <shawkins@redhat.com> (cherry picked from commit 9e6e9e3)
closes: #40099
Signed-off-by: Steve Hawkins shawkins@redhat.com
(cherry picked from commit eb96b4a)