aLIGO LHO Logbook

H1 CDS

david.barker@LIGO.ORG - posted 13:51, Wednesday 07 August 2019 - last comment - 18:38, Wednesday 07 August 2019(51105)

h1build rebooted to clear full /tmp

My daily model builds were reporting errors which I tracked down to a full /var/log file system (/tmp is linked to this). While the IFO is down, I have rebooted h1build (it had been running for 106 days).

Reminder to self: h1cdsrfm is also a non-frontend diskless machine, and it too will fill its /var/log filesystem with a large messages file after several months.

I think Keith put in a solution for this at LLO.

Comments related to this report

keith.thorne@LIGO.ORG - 18:38, Wednesday 07 August 2019 (51113)

Link

The solution is not that sophisticated
 - On Tuesday, run /etc/check_all_var_log.sh on boot server
 - for any front-end over 21,000 (20%)
 ssh controls@
 cd /var/log
 sudo mv messages messages_old
 sudo /etc/init.d/syslog_ng restart
 sudo rm messages_old
 exit