Learn more about these different git repos.
Other Git URLs
bf59b47
@@ -343,7 +343,7 @@
'subvariant': 'Container_Toolbox',
'target': global_target,
'bundle_name_format': '%N-%v-%I.%A.%T',
- 'failable': [''],
+ 'failable': ['s390x', 'ppc64le'],
},
{
'kiwi_profile': 'FEX-RootFS',
These are not release-blocking arches. Nothing on them should fail the compose. See https://pagure.io/releng/failed-composes/issue/7711#comment-951258
Signed-off-by: Adam Williamson awilliam@redhat.com
Build succeeded. https://fedora.softwarefactory-project.io/zuul/buildset/6e6017fb8a9a4cb4a288d6dd65d402f0
:thumbsup:
I suppose. I recall when we added toolbox that maintainers wanted everything to be failable, but that may have well been before we could build s390x/ppc64le containers (ie, back in the OSBS days).
CC: @rishi
Pull-Request has been merged by kevin
The decision about what's failable is more of a releng decision than an individual component maintainer decision, I'd say, as it affects everybody. I'm pretty sure we've always held that deliverables for non-release-blocking arches must be failable (i.e. must not fail the compose).
Yeah, really a fesco decision... but there was a change to make toolbox release blocking and neither in the change or the discussion I could find was a mention of what arches it applied to. ;(
Well, only x86_64 and aarch64 were added to https://docs.fedoraproject.org/en-US/releases/f41/blocking/ , so this matches that.
I didn't realize that we are now building OCI images for ppc64le and s390x. That's fantastic. Just out of curiosity, is that a consequence of using Kiwi or something else?
ppc64le
s390x
When I had submitted the ToolbxReleaseBlocker Change for Fedora 39, I was aware of only the aarch64 and x86_64 images, and my intention was to make Toolbx release-blocking on only those architectures that are release-blocking for Fedora in general. So, I won't complain about letting the Toolbx image fail on ppc64le and s390x. :)
aarch64
x86_64
I am happy to help fix the failures, though.
Yeah, the old OSBS system only had x86_64 and aarch64 clusters (you had to make a openshift cluster for each arch).
These are not release-blocking arches. Nothing on them should
fail the compose. See
https://pagure.io/releng/failed-composes/issue/7711#comment-951258
Signed-off-by: Adam Williamson awilliam@redhat.com