csfs.cs.helsinki.fi file server hung. Twice.

09.05.2014 - 03:00 - 09:50
Kuvaus: 

/cs/group file system had completely filled during the night, most probably because of some nightly cronjobs run by research groups. This triggered some Linux kernel bug and csfs file server hung completely. After csfs was brought up again at about 9:00 (the restart took longer than 60 minutes), the same cron jobs were still running. So csfs hung a second time. 

Access to group directories will now be removed from Ukko cluster. Sorry about that, but the current situation where most of department's IT systems stop working where there is too much NFS load from Ukko Cluster is not acceptable any more. At some point in future access to csfs from Ukko cluster will be completely removed. 

09.05.2014 - 10:38 Petri Kutvonen
09.05.2014 - 10:04 Jani Jaakkola