[Debian-l10n-devel] Churro down....

[Felipe Augusto van de Wiel (faw)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

On 19-03-2012 02:44, Christian PERRIER wrote:
> Quoting Martijn van Oosterhout (kleptog at gmail.com):
> 
>> Well, I can guarantee that what's running now is not what's in SVN.
>> 
>> That's mainly because I've forgotten how to commit things to SVN anyway. Then there's the question of whether we want to use the existing system or the
>> new one...
>> 
>> Where would stuff be moved to anyway?
> 
> 
> We have two options:
> 
> - i18n.debian.org, prepared by Felipe and Nicolas during last Debconf, planned to be DSA-administered. I'm not sure how well it is ready now as I don't
> exactly know the DSA requirements (Felipe does but has been very busy  since Debconf 11...too much to work on this).

In fact, DSA doesn't make packages mandatory, they even prefer if we
are able to install it manually (or the packages need to be available
thru the archive).

I'm already following -l10n-devel closer to try to help with this.
As we spoke on IRC, the ideal is that we try to schedule a Sprint
to fix things and improve the whole scenario.

> 
> - "skumbanan" a machine generously lent by a sewdish University (IIRC) with Martin Bagge as relay (local relay). The machine is ready to host us, I have
> root access and so does Martin (brother- on #debian-i18n on irc.debian.org).
> 
> skumbanan would be a temporary solution to deal with the fact that we can't even reliably reboot churro nowadays, because César's access to it is only
> indirect. The original idea was to move service from churro to skumbana as they are, without polishing things.

We agreed that once churro is back, brother would sync everyhting
out of churro, including PostgreSQL dumps and scripts.  That way
we can let churro die in peace.

Even if this means stoping a service for a while, it would be
better than lost several weeks/months of l10n teams' work.


> I planned to work on this during the March 4-11 week (where I had plenty of hacking time at home) but some unplanned events (related, indeed, to churro)
> forced me to work on something else.
> 
> If we can get churro back, I think the best is still moving thing to skumbanan as it is:
> 
> - setup the same accounts and debian-i18n-robots role

Considering what happened with Pootle, we should start from
zero and *force* people to work via SCM.  Having direct
access to the deploy makes things a pain when we need to
work on somebody's else scripts/code.


> - first setup a Debian mirror of sources as we have on churro - copy all scripts that run status pages (in /etc/cron.d/debianii18n-robots)  and cron them
> to see how things are working - move ~ddtp content - setup apache to run the DDT* things - move i18n.debian.net there

As I said, my suggestion it to make a hard move, even if
it breaks stuff.   We can then clean the house and start
moving whatever is needed to one place or another, let's
make the necessary announcements and avoid wasting people
time/work with a lost hard disk (or security issues).

Once we get things documentes and stable on brother's
machine or DSA machine, we start reactivating services
and announcing it. :)

My $0.02.

Kind regards,
- -- 
Felipe Augusto van de Wiel (faw) <faw at funlabs.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9oFfoACgkQCjAO0JDlykbSTwCg0ezWl7pkw44rXbxETtn1meia
Fy4AnAjNvO2U42PwZc3kurHabB3tB7h7
=Kbtc
-----END PGP SIGNATURE-----