[Mageia-sysadm] Toys&Tools for Mageia Infrastructure

[Kamil Rytarowski]

W dniu 03.01.2012 02:35, nicolas vigier pisze:
> On Tue, 03 Jan 2012, Kamil Rytarowski wrote:
>> Is it possible to move in a long-term, ready or working (certain) pages
>> from the mageia.org.pl server to a central one? This is a server
>> administrated by a friend of mine, that ships a few web-pages and it's
>> short on hardware and won't run everything.
> I think we can host those statistic scripts on Mageia servers. Maybe as
> part of http://dashboard.mageia.org/ ?
>
> Source code can be hosted on svn.mageia.org/svn/soft/[something], and be
> packaged in Mageia, then we can install the package on Mageia servers.
>
Great! I will contact this mailing-list on it later.
>>>> So then I have two primary questions:
>>>> - How to you find that, is it possible to (say daily, weekly) synchronize
>>>> Mageia SVN (only Cauldron tree?)? What about fetching svn:log data? It's a
>>>> few GB of data and a lot of files, directories
>>> sysadmins?
>> It would be excellent to synchronize the current SVN tree with rsync. Could
>> the sysadmins allow to do this please? Then keeping fresh list of all
>> .specs would be incomparably more friendly, including for the Mageia
>> server.
> You can synchronize a copy of the SVN using svnsync. With something like
> this :
> $ svnadmin create /path/to/local/repo
> $ svnsync init file:///path/to/local/repo svn://svn.mageia.org/svn/packages
>
> And something like this in cron :
> $ svnsync synchronize file:///path/to/local/repo
>
> I don't think svnsync takes a lot of resources on the server (except
> the first time when everything has to be downloaded), so if that's
> useful it shouldn't be a problem to run it every 15 or 30 minutes.
Thank you for the hint. I'm testing it locally right now... and for now 
it doesn't seem to be the ideal solution.

I prefer to synchronize just the needed data - to keep only the newest 
version of the ~10,000 SPEC files from the Cauldron directory-tree. This 
will be just ~30 MB of data. Mirroring the whole SVN with its history 
isn't the best solution (many gigabytes of data!) in my opinion. It 
would be doable, but I plan to move the SVN-searcher to a central 
server.. and then mirroring the whole tree isn't nice (disk usage!) if 
the script is going to be hosted on a central server.

The extracting of svn:log entries is doable only once in the beginning - 
and I prefer to annoy the SVN-server remotely to do it.

What do you think, is it possible to connect somehow RSync (or 
equivalent technology) to the directory with the freshest SVN? With 
rsync it's possible to simply skip certain directories, include the 
needed one files - with svn I haven't succeed to do this.. except to 
make single svn-sync instances for each file separately.
> You can then use your local svn mirror to get all the infos you need.
>
>> BTW. Please also fix this bug https://bugs.mageia.org/show_bug.cgi?id=2839
>> or please tell me where is the source-code of this component.
> This should now be fixed.
Thank you!