[Vagnekman-users] Re: vagn: Service Stop 0800 CET, 2010-01-27

[Per Lundqvist]

Unfortunately vagn will not be up again this evening. Due hardware
problems, our work will probably continue at least the whole tomorrow
morning.

We will notify you when it is available again.

/Vagn Admin

On Tue, 19 Jan 2010, Per Lundqvist wrote:

> Due to a planned downtime affecting Accumulus (see forwarded mail
> below), there will be a full service stop on Vagn at:
> 
>    0800 CET Wednesday January 27
> 
> During this time no jobs will be able to run and no users will be
> logged in. During the service stop we will perform OS security
> updates.
> 
> We will notify you by email when vagn is available again.
> 
> /Vagn Admin
> 
> ---------- Forwarded message ----------
> Date: Mon, 18 Jan 2010 09:39:23 +0100
> From: Kent Engström <kent at nsc.liu.se>
> To: gimle-users at lists.nsc.liu.se, bore-users at lists.nsc.liu.se
> Cc: accumulus-admin at nsc.liu.se, bore-admin at nsc.liu.se
> Subject: Planned downtime January 26-27 affecting Gimle, Tornado, Vagn, Bore
> 
> Planned downtime schedule for January 26-27
> Affecting Gimle, Tornado, Vagn, Bore
> 
> We have two planned jobs that will require downtime in the near
> future:
> 
> - Preparing the Gimle Nehalem extension for the addition of
>   24 new nodes.
> 
> - Preparing the Accumulus storage system for the addition of
>   24 new disk servers (currently undergoing acceptance
>   testing).
> 
> 
> We will combine the two downtimes to reduce the user impact.  The
> current planning is:
> 
> TUESDAY, JANUARY 26: Starting in the morning, downtime for the Nehalem
> partition on Gimle while our vendor carries out installation work and
> recabling to make room for the additional 24 nodes. The Nehalem
> partition continues to be down until the end of the stop on Wednesday.
> 
> WEDNESDAY, JANUARY 27: At noon, downtime starts for all of Gimle and
> also for the other Accumulus clients Tornado and Vagn. For a short
> time in the beginning, Bore too will be down (to be scheduled
> between runs). In the evening, the downtime ends for all systems.
> 
> Following these preparations, there will not have to be any downtime
> when the new Nehalem nodes and the new disk servers are brought
> on-line.
> 
> 

-- 
Per Lundqvist

National Supercomputer Centre
Linköping University, Sweden

http://www.nsc.liu.se