On Thu, 22 Sep 2011 20:07:39 +1200, schmitz wrote:
We might be able to trigger a reboot whenever the kernel throws a
panic() perhaps, that might cover most of the common cases. If the
system comes up without working ethernet, something like
ping -q -c 30 -w 60 gateway || reboot
run from a cron job after the network has been brought up might be
used to catch this. Booting back to a known good kernel after this
should be good enough to get the next test kernel installed and
In one of my prior companies we used this one as a hardware watchdog as
it was comparable cheap:
As far as I know you could configure two IP adresse the power switch
ping and define when it should reset the power.
Additionally you have the possibility to switch power via the
webinterface at any time...
But anyway: isn't it planned to move the machines back to NMMNs office
when they are running again? NMMN has some power switches as well and
they were willing to hook the machines up to one of those...
Ciao... // Fon: 0381-2744150
. Ingo \X/ http://blog.windfluechter.net
gpg pubkey: http://www.juergensmann.de/ij_public_key.