Fly crashed
April 9, 2013by Wm. Josiah Erikson (wjens)
Flashing caps lock and Scroll lock, no signal to VGA… nothing weird in the nagios graphs to indicate why
This is where I will let everybody know what’s up with fly, the CS cluster in ASH 130
Flashing caps lock and Scroll lock, no signal to VGA… nothing weird in the nagios graphs to indicate why
You must be logged in to post a comment.
April 9th, 2013 at 8:15 am
Rebooted just fine. Doesn’t give me a warm and fuzzy feeling….
April 9th, 2013 at 9:21 am
Well, /var/ was full. Could have something to do with it. It certainly prevented mysqld from running. Cleared out some cached RPMs and rebooted again – we should clear out /var/spool/tractor as well. Next time I reinstall fly, remember not to make /var a separate partition – SO DUMB.
April 9th, 2013 at 9:25 am
Also, note for the future: /var/ being full and mysql not running broke the check_nodes check (as well as the rocks command in general). Now, why is homedir ownership broken??
April 9th, 2013 at 2:53 pm
Turns out that it was about the hostname of the head node, which didn’t match the nodes. It has to be .local in both cases – it was set to fly.hampshire.edu on the head node, which is slightly mysterious, since /etc/hostname contained “fly.local” and nothing else.
April 9th, 2013 at 2:53 pm
Also, restarting NFS doesn’t restart rpc.idmapd – there’s a separate init script for it – /etc/init.d/rpcidmapd