Re: Debian packages containing large files
Hey guys,
Thanks for all the responses. Yes mysql replication is one of
things that we do today, however the master already has a lot of
slaves. I suppose we can setup relay slave replication but another
goal is to have a subset of consistent data so that it is more
manageable (smaller 4-5 GB rather than couple hundred GB). Perhaps, I
can play around with the source in mlib.c/libdpkg and see what's going
on? Anybody have experience doing this?
Thanks,
Sammy
On Wed, Oct 15, 2008 at 12:52 PM, Sandro Tosi <matrixhasu@gmail.com> wrote:
> Hello Sammy,
>
> On Wed, Oct 15, 2008 at 19:52, Sammy Yu <temijun@gmail.com> wrote:
>> Hey Basile,
>> Ah okay let me clarify. This package is not for public consumption
>> and would not be part of debian distribution. I'm skimming a subset
>> (percentage based) of our live production database which goes through
>> some process which generates a debian package in the end. The debian
>> package is then installed in several different test machines. This
>> process could happen daily so it doesn't make sense for the test
>> machines to be filling their local mysql install with the large
>> mysqldump which takes 6-7 hours. I recently raised the percentage and
>> noticed that the debian packages will no longer install. I'm mainly
>> looking for help in explaining why a large file in a debian package
>> would cause the install process to fail. Thanks!
>
> (Let's try to move the problem back where it belongs) did you consider
> mysql replication capabilities[1]? The configuration could be a single
> master server and many slaves (the tests server), so that the
> canonical information is stored in the current production database and
> the local copies are updated using mysql functionalities.
>
> Kindly,
> Sandro
>
> [1] http://dev.mysql.com/doc/refman/5.1/en/replication.html
>
> --
> Sandro Tosi (aka morph, Morpheus, matrixhasu)
> My website: http://matrixhasu.altervista.org/
> Me at Debian: http://wiki.debian.org/SandroTosi
>
Reply to: