[SAC] Primary Server Outlook (was: Motion: Add Alan Boudreault as SAC member and Primary Admin)

[Howard Butler]

On Oct 13, 2009, at 8:15 AM, Martin Spott wrote:

> On Tue, Oct 13, 2009 at 08:10:03AM -0500, Howard Butler wrote:
>
>> Also, Alan's experience with Xen is going to come in handy for us  
>> pretty
>> soon ;)
>
> That's interesting news. Are the Primary Servers going to get moved ?

I think we've grown to the point that it would be desirable for us to  
start segregating things by "machine."  I have been maneuvering to try  
and get the SAC at least two more servers with significant RAM that we  
can run Xen or something like it to start segregating our services and  
prepare for our move off of the very expensive Peer1 hosting.  I'm  
also concerned that OSGeo's impending budgetary realities are going to  
mean SAC's going to have to run in low-power mode fairly soon too.   
Whether we host such machines at OSL or TeleScience makes little  
difference for me, but I think we need to have the following  
requirements met:

1) Remote power management so an admin can hard reboot the machine  
over the internet
2) Remote KVM would be nice
3) Enough RAM/cores to comfortably run four Xen instances on each box

Once we did that, I would like to see us do away with the primary  
admin concept and provide sudo shell to all of the Xen instances for  
every SAC member.  I think the main reason the concept exists is  
because we're too scared to have people rooting around on osgeo1 (I'm  
scared myself, frankly).  Informal service responsibility would also  
need to be formalized a bit more in this scenario, but I think it  
could "Critical" services that we have been running on the aging  
blades should be pulled onto the new instances, while leaving lower  
priority and/or high bandwidth ones like the buildbot or osgeo4w alone.

I was confident that I was going to get two servers that met my  
criteria donated to us, but it is looking more and more like that is  
going to fall through.  My next hope is to either get an earmarked  
contribution or find some room in OSGeo's general budget to accomplish  
this.

Anyone else have any ideas for our future direction?  I don't think  
we're so sustainable in our current situation, though we can continue  
to kick the can down the road for a little while longer...

Howard