[Tornado-users] tornado: Service Stop

[Per Lundqvist]

The service stop on Tornado is now over. Due to an unfortunate
accident all jobs waiting in the batch queuing system was started, and
we had to kill all those jobs. Here is a list of the jobs that were
killed:

   94003              sm_wyser    Running    16     9:24:15  Wed Oct 29 09:00:11
   94002              sm_jenbr    Running    16     9:24:43  Wed Oct 29 09:00:39
   93986              sm_aulle    Running     1  1:01:14:05  Wed Oct 29 09:00:01
   93987              sm_aulle    Running     1  1:01:14:05  Wed Oct 29 09:00:01
   93869                x_sema    Running     4  1:23:24:22  Wed Oct 29 09:00:18
   93845              sm_jenbr    Running    16  3:07:24:39  Wed Oct 29 09:00:35
   
Our sincerest apologies for this, thanks for your patience


On Thu, 23 Oct 2008, Per Lundqvist wrote:

> Please note: the service stop is on Wednesday Oct 29, not on Thursday.
> 
> /Per 
> 
> On Wed, 22 Oct 2008, Per Lundqvist wrote:
> 
> > Please take notice: We will take down Tornado for a full service stop on 
> > Thu Oct 30, 0800 CET:
> > 
> > PLAN:
> >  . OS upgrade to latest CentOS 4
> >  . latest Lustre client (1.6.5.1)
> >  . change of Infiniband stack from IBGD 1.8.2 -> OFED 1.2.5.5 (same as
> >    on Dunder)
> >  . latest SCALI MPI software (5.6.4-59151)
> >  . firmware upgrade of HCA-cards
> >  . firmware upgrade of BMC controllers
> >  . BIOS upgrade on all machines
> >  . maintenance work on system server
> > 
> > 
> > For you this means:
> >  1. No user will be logged in during the stop.
> >  2. No jobs will be able to run from 0800 - there is a reservation
> >     within the batch queuing system preventing this (visible at:
> >     http://www.nsc.liu.se/cgi-bin/tornadostatus).
> > 
> >     Please notice that jobs submitted before this reservation, with a 
> >     walltime larger than the available time left until the reservation, 
> >     will be scheduled to run after wards.
> > 
> > 
> > The downtime is estimated to be until 1800 - you will be notified by mail 
> > when the stop is over.
> > 
> > /Per
> > 
> > 
> 
> 

-- 
Per Lundqvist

National Supercomputer Centre
Linköping University, Sweden

http://www.nsc.liu.se