Andrew Sawyers wrote:
On 10/9/06 4:47 PM, "Jim Fulton" <[EMAIL PROTECTED]> wrote:

I think the best thing to do at this point is to declare the current implementation a failure and move on.
INSHO it has been known for quite sometime, but people are still trying to
make progress in the only environment we can.


I'll repeat that our commercial cluster is not the right place to host  We can't really give volunteers the access to do what they need to
Please explain why?  There is no commercial customers on the boxes
services are on, outside of the load balancer and squid, which are not
necessary to make progress until things can be moved.

Because to do what people want to do, they think they need full sudo.
I'm not sure this is necessary.  For example, it's possible that
it would be enough for many interesting tasks to give people
the ability to sudo to the user zope.  We could do that.

Giving root access to those boxes is not a risk that we're willing
to take.

What's it going to take to move stuff?  If someone anti's up a server, can
we move things immediately so this isn't an issue?

Yes, and we may have a volunteer.  We're talking to someone now.
It will, of course, take more than a server.  The existing app
servers could probably replaced with a single modern server, although
we probably want two for redundancy.  There's a storage server, and
an ldap server.  There are squid caches and load balancers.  Of course,
we don't have to reproduce exactly the same infrastructure. Then there's
network bandwidth, system administration, etc..


Jim Fulton           mailto:[EMAIL PROTECTED]       Python Powered!
CTO                  (540) 361-1714  
Zope Corporation
Zope-web maillist  -

Reply via email to