[Systems] Load on treehouse

[David Farning]

On Sun, Aug 29, 2010 at 1:10 PM, Bernie Innocenti <bernie at codewiz.org> wrote:
> El Thu, 19-08-2010 a las 11:10 -0400, Stefan Unterhauser escribió:
>> bernie usr not monitoring ...
>>
>> and right now the main problem is that bilbo maltose dextrose openlesson and ...
>> are doing backup at the same time which really kills everything which
>> has to do with speed because our backups cost so much cpu ...
>
> Can we move these build slaves to housetree on an IPv6 subnet? SMparrish
> and rgs expressed interest in working on this.
>
> Silbe and I have experience doing it on bender.
>
>
>> by the way cpu usage score on treehouse:
>> 1) aslo-web
>> 2) lightwave
>
> I figured out why lightwave causes high load. From time to time, sks
> wakes up and trashes the disk for a few minutes. Sascha, why is that?
> What can we do to speed it up? Would defragmenting /var/lib/sks help?
>
> We should also look at setting up a second web slave for aslo. Either on
> housetree, or on another fsf machine (when I'll be back in Boston).

Now that the Servers are up at RIT we will be moving ASLO out of Gnaps....

Alsroot is mentoring a team of webdevelopers to update ASLO to the
latest version of AMO.  Upstream has migrated from cake-php to Django.
 The code is an order of magnitude cleaner:)

That should give you a lot more flexibility to move the rest of the
infrastructure to Gnaps, bender, and housetree.

david