[Tetralith-users] Tetralith login node downtime December 12th and reminder of Tetralith upgrade

[Mats Kronberg]

Dear Tetralith users,

First I would like to remind you of the ongoing Tetralith operating
system upgrade and that you will need to move your jobs to the new
operating system before January 8th. See
https://www.nsc.liu.se/support/systems/tetralith-os-upgrade/ for more
information.

As part of the operating system upgrade there will be a short downtime
for the login nodes. The date for this has now been decided: December
12th from 08:00. Expected duration: 1-4h.

Running and queued jobs should not be affected.

During this downtime, we will switch around some things to make EL9
(Rocky Linux 9) the new default.

Changes you need to be aware of:

- After the downtime, if you login to tetralith.nsc.liu.se, you will
  end up on the EL9 login node (tetralith2). Note: you can still login
  to tetralith-el7 or tetralith-el9 to manually choose which part of
  the system to use.

- During the downtime, the "devel" reservation (EL7/CentOS 7 nodes
  reserved for test and development) will be removed. As we are by then
  halfway through the migration to EL9, there should be little need for
  test and development of new CentOS 7 software. You can of course still
  use regular compute nodes for CentOS 7 test and development if you
  wish. Note: there are nodes reserved for test and development on the
  EL9 side, but the reservation is named "now".

Changes that you might notice but should not require any action on
your part:

- EL7 jobs will now run in a reservation (named "el7"). EL9 jobs will
  no longer run in a reservation.

- NSC will modify queued jobs so they are started in the correct
  reservation after the downtime.


-- 
Mats Kronberg
National Supercomputer Centre (NSC)