2015-07-22 @ 21:00 - Taskotron Outage and qadevel Migration

[Tim Flink]

# Fedora QA Devel Meeting
# Date: 2015-07-22
# Time: 21:00 UTC
(https://fedoraproject.org/wiki/Infrastructure/UTCHowto)
# Length: Approximately 4 hours

There is an outage scheduled for tomorrow, so all of the Taskotron
deployments will be down during that time.

I'm also planning to migrate qadevel to a new home - there will be no
user-facing changes at this time. We're planning to make some
significant changes to qadevel in the near future but are waiting on
some backend changes before we actually do so.

Tim

[Tim Flink]

On Tue, 21 Jul 2015 16:21:26 -0600
Tim Flink  wrote:
> # Fedora QA Devel Meeting
> # Date: 2015-07-22
> # Time: 21:00 UTC
> (https://fedoraproject.org/wiki/Infrastructure/UTCHowto)
> # Length: Approximately 4 hours
>
> There is an outage scheduled for tomorrow, so all of the Taskotron
> deployments will be down during that time.
>
> I'm also planning to migrate qadevel to a new home - there will be no
> user-facing changes at this time. We're planning to make some
> significant changes to qadevel in the near future but are waiting on
> some backend changes before we actually do so.

It took significantly longer than I was hoping but Taskotron
dev/stg/prod has been updated and qadevel has been migrated to a new
machine.

There should be no big changes here, everything user-facing should be
business as usual.

Please let me know if you notice something not quite working right.

Tim

[Kamil Paral]

> It took significantly longer than I was hoping but Taskotron
> dev/stg/prod has been updated and qadevel has been migrated to a new
> machine.
>
> There should be no big changes here, everything user-facing should be
> business as usual.
>
> Please let me know if you notice something not quite working right.

The Phabricator emails have a new sender address . Not sure it's intentional. If it is, would  be better?

None of our repos are getting synced, they report errors like these:
Initialization Error	Pull of "LTRN" failed: git fetch failed with error #255: stdout: stderr:error: cannot open FETCH_HEAD: Permission denied

I think it's the long standing issue where newly created repos can't create directories and we need to do it manually. I guess it will be some kind of directory permission issue.

[Tim Flink]

On Thu, 23 Jul 2015 05:29:17 -0400 (EDT)
Kamil Paral  wrote:
> > It took significantly longer than I was hoping but Taskotron
> > dev/stg/prod has been updated and qadevel has been migrated to a new
> > machine.
> >
> > There should be no big changes here, everything user-facing should
> > be business as usual.
> >
> > Please let me know if you notice something not quite working right.
>
> The Phabricator emails have a new sender address
> . Not sure it's intentional. If it is,
> would  be better?

We can try it, I suppose. I was mostly trying to match the domain of
the SMTP server that we're using now that we're not sending emails
directly from qadevel.

Either [e-mail] or [e-mail]
would work since we don't support ticket conversations via email and I
don't really plan to change that.
> None of our repos are getting synced, they report errors like these:
> Initialization Error	Pull of "LTRN" failed: git fetch failed
> with error #255: stdout: stderr:error: cannot open FETCH_HEAD:
> Permission denied
>
> I think it's the long standing issue where newly created repos can't
> create directories and we need to do it manually. I guess it will be
> some kind of directory permission issue.

I think this has been fixed - one of the backed changes that I made
changed the user that the phabricator daemons were running as and I
forgot to change the file permissions on the cached git repos.

If there are still problems where newly created repos can't create
directories, I'll look into it - that shouldn't happen.

Tim

[Jan Sedlak]

I've also noticed that phab.qadevel.cloud.fedoraproject.org stopped
working via HTTP - now it works only through HTTPS, so if anybody
thinks that Phabricator stopped working for him, first check that
you are using correct protocol.

Jan

[Tim Flink]

On Thu, 23 Jul 2015 08:08:12 -0400 (EDT)
Jan Sedlak  wrote:
> I've also noticed that phab.qadevel.cloud.fedoraproject.org stopped
> working via HTTP - now it works only through HTTPS, so if anybody
> thinks that Phabricator stopped working for him, first check that
> you are using correct protocol.

Yeah, port 80 isn't open to the new machine. If things go the way that
I'm hoping, this should be a temporary problem, though. If things don't
end up going the way that I'm hoping, we can get port 80 opened so that
we can do redirects to https.

Tim