FBR: open heart surgery on fedimg01
by Dusty Mabe
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On Sunday April 22nd we identified an issue with the describe-conversion-tasks
AWS API that was causing fedimg to fail when uploading images. I worked with
David Duncan from AWS (and Fedora contributor) to raise the issue with AWS
and get it resolved. The fix did not land in time for the RC compose that ran
last night so in order to publish images we need to:
- - run fedimg and let it fail
- - identify the volume that was created
- - hardcode that volume ID into the fedimg code
- - run fedimg a 2nd time to let it complete
I'd like permission to perform these actions, which I'm calling open heart surgery
so that we can get AMIs out the door for F28.
Dusty
-----BEGIN PGP SIGNATURE-----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=XQf9
-----END PGP SIGNATURE-----
6 years
[FBR] fix the alternativearch git hook
by Pierre-Yves Chibon
Good Morning Everyone,
While testing something in dist-git two days ago, we found out that there is a
piece of code missing to the alternativearch git hook.
Here is the error we saw:
Counting objects: 3, done.
Delta compression using up to 4 threads.
Compressing objects: 100% (3/3), done.
Writing objects: 100% (3/3), 384 bytes | 384.00 KiB/s, done.
Total 3 (delta 2), reused 0 (delta 0)
remote: sh: /usr/lib/python2.7/site-packages/pagure/hooks/files/fedmsg_hook.py: No such file or directory
remote: Emitting a message to the fedmsg bus.
remote: * Publishing information for 1 commits
remote: Traceback (most recent call last):
remote: File "/usr/share/git-core/post-receive-alternativearch", line 206, in <module>
remote: run_as_post_receive_hook()
remote: File "/usr/share/git-core/post-receive-alternativearch", line 163, in run_as_post_receive_hook
remote: new_commits_list = get_revs_between(oldrev, newrev, abspath, refname)
remote: File "/usr/share/git-core/post-receive-alternativearch", line 96, in get_revs_between
remote: head = get_default_branch(abspath)
remote: NameError: global name 'get_default_branch' is not defined
remote: Sending to redis to send commit notification emails
remote: * Publishing information for 1 commits
remote:
remote: Create a pull-request for f27
remote: https://src.stg.fedoraproject.org/rpms/guake/diff/master..f27
remote:
To ssh://pkgs.stg.fedoraproject.org/rpms/guake
* [new branch] f27 -> f27
And here is the patch I would like to apply to fix it:
diff --git a/roles/git/hooks/files/post-receive-alternativearch b/roles/git/hooks/files/post-receive-alternativearch
index 7bc4a9a6a..8d93137e4 100755
--- a/roles/git/hooks/files/post-receive-alternativearch
+++ b/roles/git/hooks/files/post-receive-alternativearch
@@ -100,6 +100,15 @@ def get_revs_between(oldrev, newrev, abspath, refname):
return read_git_lines(cmd, abspath)
+def get_default_branch(abspath):
+ """ Return the default branch of a repo. """
+ cmd = ['rev-parse', '--abbrev-ref', 'HEAD']
+ out = read_git_lines(cmd, abspath)
+ if out:
+ return out[0]
+ else:
+ return 'master'
+
def send_email(text, subject, to_mail): # pragma: no cover
''' Send an email with the specified information.
Thoughts?
Pierre
6 years
Fix cron job fixing/adjusting the git hooks in dist-git
by Pierre-Yves Chibon
Good Morning Everyone,
I re-wrote last week the script fixing the git hooks in dist-git, but its work
is getting overridden by the old script running in a cron (that Kevin fixed this
week-end).
So I would like to replace the cron running the old script with a cron running
the new script:
Here is the diff:
diff --git a/ roles/gitolite/check_fedmsg_hooks/tasks/main.yml b/ roles/gitolite/check_fedmsg_hooks/tasks/main.yml
index d357a7d83..bea193af1 100644
--- a/ roles/gitolite/check_fedmsg_hooks/tasks/main.yml
+++ b/ roles/gitolite/check_fedmsg_hooks/tasks/main.yml
@@ -8,7 +8,7 @@
minute=10
hour="0,12"
user=pagure
- job="/usr/local/bin/git-check-perms /srv/git/repositories --check=fedmsg-hook -f"
+ job="/usr/local/bin/distgit_check_hook.py"
tags:
- git
- gitolite
Note, we could make this run more often as it runs in a few minutes (the forks
being the slowest namespace, the other namespace run under a minute).
Thoughts?
Pierre
6 years
FBR: Make fpaste.org cert use letsencrypt
by Ricky Elrod
Right now fpaste.org uses the getfedora.org cert which means it errors
for everyone when they access it over https. I have generated a cert
for it on certgetter01 and would like to run the proxy playbook to
sync it out.
I'm going to try to do so with a tag I added (-t fpaste.org) so that
it only changes that one site, but I'm not sure if roles can be tagged
that way. If they can't, I'll have to do a full proxy playbook run.
+1s?
6 years
FBR: bodhi-3.6.1
by Randy Barlow
Gr33tings, people of Fedora Infra,
I would like to deploy Bodhi 3.6.1[0,1] to production, which should
address some of the stability issues we experienced with 3.6.0. It will
also resolve something that's made me uncomfortable over the past week -
Bodhi's backends have been running 3.6.0 while the front ends have been
running 3.5.2. In theory this should be fine, but it makes me nervous.
In addition to the changes upstream, I've made some downstream and
deployment configuration changes in this release:
* The Rawhide and f27-infra packages now split the composer out into its
own subpackage[2]. This greatly reduces the size of the Bodhi web
container (cutting dependencies about in half due to not needing
Pungi).
* I switched Bodhi to use an in-memory cache for the home page instead
of a filesystem cache[3]. Patrick and I believe that the filesystem
cache caused some of the OpenShift instability due to the server
processes waiting on file locks. The real patch for this will involve
removing some "if staging" statements from the cited commit, but I
didn't want to send *that* patch because it would be less clear than
the one I cited as to what the real change is.
I believe that this release will be stable in OpenShift, though it's
hard to be 100% sure since the majority of the issues only happen under
production load.
[0] https://bodhi.stg.fedoraproject.org/docs/user/release_notes.html
[1] https://github.com/fedora-infra/bodhi/compare/3.6.0...3.6.1[2]
https://src.fedoraproject.org/rpms/bodhi/c/6b63e0dc5d340cdac80ca3fcc2cf7e...
https://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=c99e...
6 years