[Systems] Status of ASLO at RIT
Stephen Jacobs
itprofjacobs at gmail.com
Wed Mar 3 06:38:16 EST 2010
hope to meet with folks On this today
Sent from my iPhone
On Feb 28, 2010, at 3:08 PM, Bernie Innocenti <bernie at codewiz.org>
wrote:
> On Sat, 2010-02-27 at 20:49 -0600, David Farning wrote:
>> Can can you fill in the specification of the three aslo servers for
>> Steven? I think you know all of the design requirements. Every thing
>> is pretty standard. The service itself will only need one public IP
>> which 'floats' between squid primary and squid secondary. Rather than
>> waste IP could we can the maintenance ips ipv6? Dogi has purchased or
>> will purchase device to remote power cycle the machines.
>>
>> I don't think we will need to worry about backing up the machines. I
>> have automated every thing so that it will be easier to configure a
>> new machine via puppet than restore a backup. The only state we need
>> to worry about are the database and the nfsfile system which will
>> exist on at least 2 machines in the cluster and will get backed up
>> separately.
>
> I think you guys know better than me, but I'll try to summarize here:
>
>
> * rack space and power for 3 x 1U servers.
>
> * 4 public IP addresses
>
> * If possible. global IPv6 addresses should be routed as well
> (otherwise we'll use 6to4)
>
> * Isolated gigabit LAN to interconnect the machines.
> (this could be done with port management in the switch or
> with a small dedicated switch).
>
> * Contacts of on-site sysadmins who could assist us in case of
> emergencies (best-effort, no obligation).
>
> * Provide physical access to Sugar Labs sysadmins for initial
> setup and future maintenance. The initial setup will be carried
> on by Stefan "dogi" Unterhauser.
>
> * Bandwidth usage is projected to peak at 15-20Mbit/s during
> morning hours (GMT+3).
>
> * We don't need space for backups as these servers won't store
> valuable data and are easy to reconstruct from scratch
>
>
> If RIT agrees to host it, at a later time we might add one more 1U
> or 2U
> server running our primary websites: wiki.sugarlabs.org,
> planet.sugarlabs.org, lists.sugarlabs.org and others. It's expected to
> use less bandwidth than ASLO, though.
>
> Thanks,
>
> --
> // Bernie Innocenti - http://codewiz.org/
> \X/ Sugar Labs - http://sugarlabs.org/
>
More information about the Systems
mailing list