Dev/tsweeney/b1.15.1 #2495

TomSweeneyRedHat · 2020-07-27T22:58:35Z

As the title says!

This was my first big cherry-pick and kit build on a release branch. I'd problems with the cherry-pick, it didn't like one of the buildahimage/Containerfiles, so hopefully I didn't mess it up much.

Per https://docs.docker.com/storage/bind-mounts/#configure-mount-consistency-for-macos this is just another way to specify the already ignored "cached" and "delegated" mount options. Signed-off-by: Marius Hoch <hoch.m00@gmail.com>

Apparently docker supports this as well, see https://docs.docker.com/storage/volumes/#use-a-read-only-volume Signed-off-by: Marius Hoch <hoch.m00@gmail.com>

rhatdan · 2020-07-28T12:06:02Z

Validate is not happy.
LGTM once test pass.

TomSweeneyRedHat · 2020-07-28T17:03:08Z

@rhatdan I see everything in yellow here. Where are you seeing an error? But then again, I'm not sure why it's yellow 12+ hours later....

TomSweeneyRedHat · 2020-07-28T17:04:09Z

OK, git weirdness, despite having refreshed my web interface, no change to the build status on the GitHub page. I did the last reply and now I'm seeing the red. I love computers....

As the title says. Signed-off-by: TomSweeneyRedHat <tsweeney@redhat.com>

TomSweeneyRedHat · 2020-07-28T17:52:47Z

@rhatdan I cherry picked a number of Commits from your note, but three don't seem to be included. Still looking, they are:
chroot: do not use setgroups if it is blocked - 2b1ad12
chroot, run: not fail on bind mounts from /sys - c43ab20
Mask over the /sys/fs/selinux in mask branch - c00b434

if setgroups is blocked to set up the user namespace, do not attempt to use it to clear the additional groups. Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>

some file systems under /sys might not be accessible to an unprivileged user. Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>

TomSweeneyRedHat · 2020-07-28T17:56:10Z

OK, now they're all there. I'm not sure why they didn't pull with the mass cherry-pick that I did to start. Fingers crossed for the testing.

TomSweeneyRedHat · 2020-07-28T23:39:36Z

@nalind @rhatdan Gating is falling down with the below, but I don't think this commit is part of this PR, and if it is, I'm not sure what magic git commands I'd need to fix it? I just checked my command history, I did a git commit --amend -s as part of the submit process. Help!

[+0067s]  * f7ff4c11 "Merge pull request #2296 from TomSweeneyRedHat/bump-1.14.8" ... FAIL
[+0067s]   - FAIL - does not have a valid DCO

nalind · 2020-07-29T13:57:12Z

The validation test that runs as part of the gating job looks like it's picking a common ancestor commit from master as the baseline. Going by contrib/cirrus/test.sh, it's using $DEST_BRANCH to direct the search for that ancestor. Changing its value in .cirrus.yml from "master" to "release-1.15" looks to me like the right change to make to fix that.

nalind · 2020-07-29T13:58:52Z

CHANGELOG.md

+    vendor golang.org/x/text@v0.3.3
+    Cirrus: Fix missing htpasswd in registry img
+    Switch scripts to use containers.conf
+    Merge pull request #2471 from vrothberg/1.15-update-x/text


This is the merge commit log for line 12. We don't need to mention it.

nalind · 2020-07-29T13:59:06Z

CHANGELOG.md

+    Switch scripts to use containers.conf
+    Merge pull request #2471 from vrothberg/1.15-update-x/text
+    Make imagebuildah.BuildOptions.Architecture/OS optional
+    Merge pull request #2483 from nalind/release-1.15-arch


This is the merge commit log for the previous line. We don't need to mention it.

TomSweeneyRedHat · 2020-07-29T20:03:49Z

@nalind, thx very much for the cirrus pointer. I've touched that up and have removed the commit titles as you requested from the changelogs and buidlah.spec.

TomSweeneyRedHat · 2020-07-29T20:12:07Z

Unfortunately still the same DCO issue for that commit....

nalind · 2020-07-29T20:51:15Z

Based on the test log's Linting & Validating from $GITVALIDATE_EPOCH to $GITVALIDATE_TIP output, contrib/cirrus/test.sh is getting the right starting point from git merge-base. The top-level Makefile's validate target doesn't override that when the script invokes make.
tests/validate/git-validation.sh appears to override it to c5546750573cdef524be7159fe837f23274e9718 if "$TRAVIS" isn't set, which looks like the rest of the problem. That default should probably be getting set to the release-1.15 branch's origin point (3b1d6ebe12445dfe84cb5932634440a20b10fc03) if "$CIRRUS_CI" isn't set.

This is required so that the mount point shows up when buildah is vendored into Podman. Signed-off-by: Daniel J Walsh <dwalsh@redhat.com> Signed-off-by: TomSweeneyRedHat <tsweeney@redhat.com>

TomSweeneyRedHat · 2020-07-29T21:46:35Z

thx again @nalind, tweaked and repushed. I also added a comment in the git-validation.sh file explaining where the id there is getting pulled from.

TomSweeneyRedHat · 2020-07-29T22:10:29Z

@nalind, @cevich Looks like I'm finally past gating, now I'm hitting an artifact problem?

Uploading 1 artifacts for /var/tmp/go/src/github.com/containers/buildah/bin/*
Failed to upload artifact file /var/tmp/go/src/github.com/containers/buildah/bin/buildah: EOF

Thoughts?

nalind · 2020-07-30T15:46:15Z

LGTM

nalind · 2020-07-30T18:11:07Z

@nalind, @cevich Looks like I'm finally past gating, now I'm hitting an artifact problem?
Uploading 1 artifacts for /var/tmp/go/src/github.com/containers/buildah/bin/*
Failed to upload artifact file /var/tmp/go/src/github.com/containers/buildah/bin/buildah: EOF
Thoughts?

The branch's .cirrus.yml includes multiple build_artifacts keys that instruct Cirrus to grab the contents of bin/*, but the top-level Makefile doesn't put binaries there. I'm not sure how a file that it failed to upload even got there.

cevich · 2020-07-30T20:22:56Z

I'm not sure how a file that it failed to upload even got there.

One of the CI scripts creates that link for this purpose. I think if there was some other primary-failure, failing to upload could be an expected secondary failure. I would just ignore it unless it persists beyond the primary (build) failure.

nalind · 2020-07-30T20:33:00Z

Ah, the binaries are written to the bin directory on the master branch, but that wasn't the case yet when release-1.15 was branched. A dangling symlink might explain that error.

TomSweeneyRedHat · 2020-07-30T20:50:53Z

We're all green here now. @cevich or @nalind, did one of you force a retest? I'm not seeing evidence in the git logs above....

nalind · 2020-07-30T20:59:18Z

I pushed the retest button a couple of times, yes.

TomSweeneyRedHat · 2020-07-30T21:27:34Z

Ah, mystery solved, thanks @nalind.

rhatdan · 2020-07-30T21:50:29Z

bors r+

2495: Dev/tsweeney/b1.15.1 r=rhatdan a=TomSweeneyRedHat As the title says! This was my first big cherry-pick and kit build on a release branch. I'd problems with the cherry-pick, it didn't like one of the buildahimage/Containerfiles, so hopefully I didn't mess it up much. Co-authored-by: Marius Hoch <hoo@online.de> Co-authored-by: Zvi "CtrlZvi" Effron <viz_skywalker+GitHub@outlook.com> Co-authored-by: Giuseppe Scrivano <gscrivan@redhat.com> Co-authored-by: Daniel J Walsh <dwalsh@redhat.com>

nalind · 2020-07-30T22:17:43Z

I'm not sure how a file that it failed to upload even got there.

One of the CI scripts creates that link for this purpose. I think if there was some other primary-failure, failing to upload could be an expected secondary failure. I would just ignore it unless it persists beyond the primary (build) failure.

If this is a script rather than something Cirrus does because we set binary_artifacts in .cirrus.yml, I can't find it; can you point me to where this is done? I got the same error in https://cirrus-ci.com/task/5209101888126976, which is for a PR that isn't on this branch. Re-running the test made it succeed as well, but I'm no longer sure there isn't a problem we need to fix.

bors · 2020-07-30T22:40:58Z

Build failed:

cirrus-ci/success

vrothberg · 2020-07-31T07:14:36Z

bors retry

2495: Dev/tsweeney/b1.15.1 r=rhatdan a=TomSweeneyRedHat As the title says! This was my first big cherry-pick and kit build on a release branch. I'd problems with the cherry-pick, it didn't like one of the buildahimage/Containerfiles, so hopefully I didn't mess it up much. Co-authored-by: Marius Hoch <hoo@online.de> Co-authored-by: Zvi "CtrlZvi" Effron <viz_skywalker+GitHub@outlook.com> Co-authored-by: Giuseppe Scrivano <gscrivan@redhat.com> Co-authored-by: Daniel J Walsh <dwalsh@redhat.com>

bors · 2020-07-31T07:54:30Z

Build failed:

cirrus-ci/success

rhatdan · 2020-07-31T11:27:17Z

bors retry
@TomSweeneyRedHat once this merges please cut a release and get this in Podman. 2.0.4
@mheon FYI

bors · 2020-07-31T12:09:44Z

Build succeeded:

cirrus-ci/success

cevich · 2020-07-31T15:21:24Z

I can't find it; can you point me to where this is done?

Oh! Hmm, it was right around here https://github.com/containers/buildah/blob/master/contrib/cirrus/build.sh#L22

but it's gone now 😕 I swore I...oh, here's what I was thinking of: 90868620d#diff-43090a914b41249c6581b485cd9a4e06R22-R25

Does that help?

nalind · 2020-07-31T18:29:30Z

I guess we removed when we started putting the binaries there in #2380? I don't know enough about the packaging system to know whether it was a requirement for making it work or if it was just a stylistic change that happened at the same time.

mariushoch added 2 commits July 27, 2020 18:47

Ignore OS X specific consistency mount option

36f8c92

Per https://docs.docker.com/storage/bind-mounts/#configure-mount-consistency-for-macos this is just another way to specify the already ignored "cached" and "delegated" mount options. Signed-off-by: Marius Hoch <hoch.m00@gmail.com>

Allow "readonly" as alias to "ro" in mount options

b56af2b

Apparently docker supports this as well, see https://docs.docker.com/storage/volumes/#use-a-read-only-volume Signed-off-by: Marius Hoch <hoch.m00@gmail.com>

Bump to v1.15.1

24c307c

As the title says. Signed-off-by: TomSweeneyRedHat <tsweeney@redhat.com>

TomSweeneyRedHat force-pushed the dev/tsweeney/b1.15.1 branch from 3c31664 to 24c307c Compare July 28, 2020 17:37

giuseppe added 2 commits July 28, 2020 13:53

chroot: do not use setgroups if it is blocked

e9ed213

if setgroups is blocked to set up the user namespace, do not attempt to use it to clear the additional groups. Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>

chroot, run: not fail on bind mounts from /sys

4c52b9e

some file systems under /sys might not be accessible to an unprivileged user. Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>

nalind reviewed Jul 29, 2020

View reviewed changes

TomSweeneyRedHat force-pushed the dev/tsweeney/b1.15.1 branch from 4e42b4a to 2fd3fc9 Compare July 29, 2020 20:02

Mask over the /sys/fs/selinux in mask branch

cd67d69

This is required so that the mount point shows up when buildah is vendored into Podman. Signed-off-by: Daniel J Walsh <dwalsh@redhat.com> Signed-off-by: TomSweeneyRedHat <tsweeney@redhat.com>

TomSweeneyRedHat force-pushed the dev/tsweeney/b1.15.1 branch from 2fd3fc9 to cd67d69 Compare July 29, 2020 21:45

bors bot merged commit c967e88 into containers:release-1.15 Jul 31, 2020

TomSweeneyRedHat deleted the dev/tsweeney/b1.15.1 branch August 14, 2020 20:37

github-actions bot added the locked - please file new issue/PR label Sep 30, 2023

github-actions bot locked as resolved and limited conversation to collaborators Sep 30, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dev/tsweeney/b1.15.1 #2495

Dev/tsweeney/b1.15.1 #2495

TomSweeneyRedHat commented Jul 27, 2020

rhatdan commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

nalind commented Jul 29, 2020 •

edited

Loading

nalind Jul 29, 2020

nalind Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

nalind commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

nalind commented Jul 30, 2020

nalind commented Jul 30, 2020

cevich commented Jul 30, 2020

nalind commented Jul 30, 2020

TomSweeneyRedHat commented Jul 30, 2020

nalind commented Jul 30, 2020

TomSweeneyRedHat commented Jul 30, 2020

rhatdan commented Jul 30, 2020

nalind commented Jul 30, 2020

bors bot commented Jul 30, 2020

vrothberg commented Jul 31, 2020

bors bot commented Jul 31, 2020

rhatdan commented Jul 31, 2020

bors bot commented Jul 31, 2020

cevich commented Jul 31, 2020

nalind commented Jul 31, 2020

Dev/tsweeney/b1.15.1 #2495

Dev/tsweeney/b1.15.1 #2495

Conversation

TomSweeneyRedHat commented Jul 27, 2020

rhatdan commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

TomSweeneyRedHat commented Jul 28, 2020

nalind commented Jul 29, 2020 • edited Loading

nalind Jul 29, 2020

Choose a reason for hiding this comment

nalind Jul 29, 2020

Choose a reason for hiding this comment

TomSweeneyRedHat commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

nalind commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

TomSweeneyRedHat commented Jul 29, 2020

nalind commented Jul 30, 2020

nalind commented Jul 30, 2020

cevich commented Jul 30, 2020

nalind commented Jul 30, 2020

TomSweeneyRedHat commented Jul 30, 2020

nalind commented Jul 30, 2020

TomSweeneyRedHat commented Jul 30, 2020

rhatdan commented Jul 30, 2020

nalind commented Jul 30, 2020

bors bot commented Jul 30, 2020

vrothberg commented Jul 31, 2020

bors bot commented Jul 31, 2020

rhatdan commented Jul 31, 2020

bors bot commented Jul 31, 2020

cevich commented Jul 31, 2020

nalind commented Jul 31, 2020

nalind commented Jul 29, 2020 •

edited

Loading