[Planetlab-users] Re: extent of slice disability

[Larry Peterson]

By the way, if you want to see whether PLC thinks your nodes
are down, take a look at

   http://monitor.planet-lab.org/

It's the "Offline" nodes that will soon be subject to slice
suspension. The "Unreachable" nodes are suffering some short-term
problems, although we need to be more diligent in moving nodes
from Unreachable to Offline.

Larry

On May 13, 2005, at 10:17 AM, Larry Peterson wrote:

> Good questions. We're still working out the details, but the
> most likely answer is that you'd get an explanation when you
> tried to ssh into your slice. I don't think we have plans to
> delete the slice from nodes, but that might be the long-term
> answer.
>
> That a site's slices are about to be suspended shouldn't come
> as a surprise. It will only happen after multiple failed attempts
> to get a response, or continued foot-dragging about some policy
> issue. Our experience is that unless users have some skin in the
> game, there's little motivation for them to keep tabs on their
> own machines and help resolve any problems that arise.
>
> Larry
>
>
> On May 13, 2005, at 9:50 AM, Emil Sit wrote:
>
>> On Thu, 12 May 2005 at 14:14 (-0400), Larry Peterson wrote:
>>> avail. We've decided to add a new last-resort mechanism, which
>>> is to disable a site's slices when its nodes are down for an
>>
>> Does "disable" mean that the slice will be deleted from nodes
>> (i.e. with data loss) or just that the slices processes will be
>> killed (but with data left alone, and consuming shared resources)?
>>
>> What error message will be displayed if an authorized user
>> tried to ssh into a disabled sliver/slice?
>>
>> -- 
>> Emil Sit / MIT: CSAIL PDOS, SIPB, OLC / http://web.mit.edu/sit/www/
>>
>
> _______________________________________________
> Users mailing list: Users at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/users
>