[Tornado-users] Reboot of Tornado login node today at 11:00 due to Out-of-Memory condition

[Johan Raber]

Dear users,

At 11:00 today the login node of Tornado will need to be rebooted due to an
OOM incident yesterday, where a user process consumed all available memory
and was consequently killed. Very often this leads to unpredictable
behaviour of the computer and necessitates a reboot.

To avoid these types of incidents on the login node, two approaches can be
warmly recommended; allocate a cluster node for interactive work or use
either of the two reserved analysis nodes (a1, a2) which can be reached
from the login node via SSH.

If you are unfamiliar with how to accomplish these things we recommend a
look at the Tornado user guide at URL
http://www.nsc.liu.se/systems/cluster/tornado/.
Should you not find the answers you need there, please contact
smhi-support at nsc.liu.se

Best regards,
NSC support