Kick-off meeting for new RackSpace-based hosts

[Karsten 'quaid' Wade]

On 01/05/2013 08:48 AM, Eyal Edri wrote:
> 
> 
> ----- Original Message -----
>> From: "Karsten 'quaid' Wade" <kwade at redhat.com>
>> To: "infra" <infra at ovirt.org>
>> Sent: Friday, January 4, 2013 10:41:45 PM
>> Subject: Re: Kick-off meeting for new RackSpace-based hosts
>>
>> On 01/04/2013 07:23 AM, Mike Burns wrote:
>>> On Thu, 2013-01-03 at 16:14 -0500, Eyal Edri wrote:
>>
>>>> Yes. please send me the call details when you have them.
>>>> I believe those servers can be used for:
>>>> 1. one or more server to be used as hypervisor to run jenkins
>>>> slave vm with multiple os's
>>>> 2. one physical server to be used for automatic testing of ovirt
>>>> (adding host,etc...)
>>>
>>> What about backups?  Can we host that somewhere there? or do we
>>> have
>>> somewhere else for storing backups?
>>
>> Right, this is part of the still-open discussion of how we want to
>> use
>> all the new hosts.
>>
>> I wasn't aware it was valuable to have a bare metal host for
>> automatic
>> testing, but that's not a problem. Could we potentially host VMs on
>> that
>> host, or does it need to be able to be stripped and replaced
>> completely
>> each time.
> 
> i don't think it will be possible to have a host to run vms and be used for automatic tests as hypervisor at the same time.
> another alternative will be to use nested hosts. 
> i'm currently in the process of testing it on f17, having bit of issue with qemu though.
> if it will prove working, we can use a nested vm instead of a bare-metal host to run automation tests. 

I'm going to take my response back to the "New hosting design" thread,
but the tl:dnr version is, maybe we need to have
random-useful-VM-hosting at Alter Way and keep the RackSpace-based hosts
for i) Jenkins slaves, ii) automated oVirt testing.

BTW, how are we running automated testing?

- Karsten

> e.
> 
>>
>> For backup, I had been thinking of something like this:
>>
>> * Run Gluster at each site so the storage pool is only pulling from
>> the
>> LAN (or localhost.)
>> * Use some type of Gluster mirroring so that each location contains a
>> backup of the other location.
>>
>> Anyway, I haven't checked with Glustexperts yet on that scheme, but I
>> think the basic idea is workable.
>>
>> In this scenario, one physical host would have 6 to 12 VMs. One VM
>> could
>> be dedicated to all sorts of cronjobs and batchwork - copying built
>> RPMs
>> from one location to another, backing up files and databases, etc.
>> This
>> host could be entirely on the private VLAN, connecting to external
>> hosts
>> with ssh (scp, rsync, tunneling, Ansible, Puppet, whatever) or a VPN.
>>
>> - Karsten
>> --
>> Karsten 'quaid' Wade, Sr. Analyst - Community Growth
>> http://TheOpenSourceWay.org  .^\  http://community.redhat.com
>> @quaid (identi.ca/twitter/IRC)  \v'  gpg: AD0E0C41
>>
>>
>> _______________________________________________
>> Infra mailing list
>> Infra at ovirt.org
>> http://lists.ovirt.org/mailman/listinfo/infra
>>


-- 
Karsten 'quaid' Wade, Sr. Analyst - Community Growth
http://TheOpenSourceWay.org  .^\  http://community.redhat.com
@quaid (identi.ca/twitter/IRC)  \v'  gpg: AD0E0C41

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 253 bytes
Desc: OpenPGP digital signature
URL: <http://lists.ovirt.org/pipermail/infra/attachments/20130106/4bf6e8bf/attachment.sig>