On Tue, 2016-06-28 at 23:04 -0700, Adam Williamson wrote:
On Tue, 2016-06-28 at 15:10 -0700, Adam Williamson wrote:
>
> This week I'm planning - if Emmanuel Seyman from the perl SIG is OK
> with it - to send out the same three builds to updates-testing for F24,
> and update staging to the new openqa and os-autoinst. If that all
> appears to go OK, I'll move on to updating prod to F24 and the new
> openqa/os-autoinst.
So I sent the update out and upgraded staging, and kinda...uh...broke
it a bit. It's forgotten what test all the existing jobs were for,
which makes things look a bit funky when you look at existing jobs. But
oh well! That's why it's staging. I might be able to fix it if we have
backups of the old database data, if not, eh, it's not a huge deal.
Future jobs will be fine.
The problem happened because database upgrade failed for some reason
and I screwed up a bit trying to re-do the upgrade manually; I know
what I did, though, so if the same thing happens when we upgrade prod I
promise not to mess it up...
Not sure yet why the initial problem happened, it didn't on my test
box. Oh, well.
Thanks to puiterwijk, the database is fixed again now and stg should be
behaving OK. I do think there's something messed up with the db upgrade
scripts upstream, though, so won't be updating prod till I've talked to
upstream about it, though I think I know how to handle it manually if
necessary now.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
http://www.happyassassin.net