[Neolith-users] FYI: Short unplanned outage on the Neolith login node (neolith1)

[Mats Kronberg]

Just FYI: the Neolith login node (neolith1 a.k.a neolith.nsc.liu.se) was 
unavailable from ~16:10 to 16:25 CEST today.


Details:

During the afternoon, we were investigating a disk performance problem 
on neolith1. At 16:10 CEST accessing the shared storage (/home, 
/nobackup/global, /software) on neolith1 went from being slow to totally 
unresponsive, making it impossible to login or to any meaningful work.

Short-term fix: at 16:25 we rebooted neolith1.

Long-term fix: We'll extract whatever information is in the logs about 
the disk problem and try to prevent it from happening again.


No running or queued jobs were affected. No filesystem data was lost. 
Everything should now be back to normal.



-- 
Mats Kronberg, Systems Expert
NSC, National Supercomputer Centre
Email: support at nsc.liu.se