test: [sig-builds][Feature:Builds] forcePull should affect pulling builder images ForcePull test case execution s2i is failing frequently in CI, see search results: https://search.svc.ci.openshift.org/?maxAge=168h&context=1&type=bug%2Bjunit&name=&maxMatches=5&maxBytes=20971520&groupBy=job&search=%5C%5Bsig-builds%5C%5D%5C%5BFeature%3ABuilds%5C%5D+forcePull+should+affect+pulling+builder+images++ForcePull+test+case+execution+s2i
This was caused by registry.redhat.io issues. It should be ok now, but it would be good to use this bug to confirm the test is passing again in recent runs.
Based on the serach.svc query no imagestream related problems have occurred since 2 days ago when the TBR problems abated. So Ben's ask in #Comment 1 I believe is addressed. That said, these forcepull tests have been skipped for quite a while, re: https://github.com/openshift/origin/blob/master/test/extended/builds/forcepull.go#L112-L134 @Adam @Ben do we think the buildah based caching stuff will re-introduced the advantages of forcePull being set to true? Or is it time to just delete these tests.
> @Adam @Ben do we think the buildah based caching stuff will re-introduced the advantages of forcePull being set to true? Or is it time to just delete these tests. i think it will reintroduce the value... my expectation is that if the "tag" exists in the local image cache, no registry check/pull will be performed, which would mean you get potentially stale content unless you do a forcepull. But it might be worth confirming what the behavior will be with the buildah folks. They may only be caching layers by sha in which case the registry would always be consulted to get the latest shas for the tag, prior to cheecking the local cache for those shas, in which case there is no value in explicitly saying "force pull"
(In reply to Ben Parees from comment #3) > > @Adam @Ben do we think the buildah based caching stuff will re-introduced the advantages of forcePull being set to true? > Or is it time to just delete these tests. > > i think it will reintroduce the value... my expectation is that if the "tag" > exists in the local image cache, no registry check/pull will be performed, > which would mean you get potentially stale content unless you do a forcepull. > > But it might be worth confirming what the behavior will be with the buildah > folks. They may only be caching layers by sha in which case the registry > would always be consulted to get the latest shas for the tag, prior to > cheecking the local cache for those shas, in which case there is no value in > explicitly saying "force pull" Good point re: Buildah details @Nalin - do you have enough context here in this BZ address whether the current trend wrt buildah caches will be able to provide value for an openshift build's controls around always pulling its builder image? If not, perhaps we can discuss in the next devex scrum.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:2409