[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: UDD migration to ullmann.debian.org

This one time, at band camp, Andreas Tille said:
> On Wed, Jun 27, 2012 at 07:33:29AM +0100, Stephen Gran wrote:
> > > I wonder whether we could now find a way to export the UDD bugs
> > > tables as I have suggested and which was confirmed by Lucas:
> > > 
> > >    http://lists.debian.org/debian-qa/2012/05/msg00023.html
> > 
> > I see the advantage of having a udd replica set somewhere,
> What I'm running is not really a replica of UDD.  I use
> blends.debian.org to test new code which is not finished and thus can
> not go to official UDD.  My point in cloning the bugs tables was that
> for the purpose of testing the effort of having a >65GB clone of the
> BTS as well does not seem to be rectified.

When I say, "a replica set", I'm speaking of the database.  And I agree,
cloning 65GB of data that is an api call away does seem wasteful.

> > if
> > latency and read performance on ullmann is problematic.  As we haven't
> > tried ullmann yet, though, it seems a little premature.  Also, it seems
> > you are still loading up alioth with random jobs that have little to do
> > with alioth's primary goal of being a software forge.  Are you having
> > difficulty finding hosting space for your pet projects?
> What you might call "random jobs"

[... stuff about a cron job ...]

No, I don't think that a cron job that looks at subproject metadata is
related to collaborative software development in Debian, or at least not
directly.  I'm not going to say stop, but I'd prefer not to encourage

Back to the topic at hand:

> > > I do see two options:
> > > 
> > >    1. Either generate the partial database dump in a cron
> > >       job and move it to some http-accessible space or

There is already a database dump accessable over http, isn't there?

> > >    2. Get a passwordless sshkey which has the only single
> > >       permission to trigger the dump and rsync the result
> > >       (=call a dedicated scipt)

If you mean a cron job that pushes from ullmann to your machine, I don't
mind how you implement it.  If you mean you want to put a passwordless
ssh key on ullmann and run a command there, I'd rather you didn't.

> > postgres has native replication support.  I don't see a need to reinvent
> > the wheel here, do you?
> No, I don't.  But I would also like to replicate some specific tables
> and I'd be really happy to learn how to do this in the most reasonable
> way.
> > > What would your prefered solution to make a a dump of a few UDD tables
> > > accessible to some other host?
> > 
> > What other host?
> blends.debian.net .

Well, we're obviously not going to replicate the database to a machine
we don't run.  Never mind, then.

Maybe if you ask lucas nicely, he can dump tables one at a time for you
so you have what you need?  At any rate, the problem of copying data
over http from a debian machine to a non-debian machine probably doesn't
need DSA to do very much, so I'll leave it here.

|   ,''`.                                            Stephen Gran |
|  : :' :                                        sgran@debian.org |
|  `. `'                        Debian user, admin, and developer |
|    `-                                     http://www.debian.org |

Attachment: signature.asc
Description: Digital signature

Reply to: