Drew Van Zandt <[EMAIL PROTECTED]> writes: > http://www.underhanded.org/papers/debian-conversion/remotedeb.html > > Modify for distro?
Heh, I remember reading this on /. when it was first announced. We recently just automated our FAI config even more... We've been having a real problem with drive reliability lately which results in the machine in question needing to be re-installed (the re-partitioning of the drives vectors around the newly found bad sectors). The problem was that this required manual intervention of someone going into the lab, hooking up a monitor/keyboard and on the console, issuing a reboot command, then waiting for the POST so we could hit F10 or whatever to force a PXE boot. However, since we have such a complex and automated FAI config, all that was missing was a way to force the systems to reboot into PXE mode (Sun hw was good for something:). Then it dawned on us that we could have the BIOS boot sequence cascade from hard drive to net. So now, with SMARTD monitoring our drives, when it detects an error, it execs a script which takes the system out of our automated-test reservation system and flags it for re-install. Then we manually issue the re-fai command, which does little more than: dd if=/dev/zero of=/dev/hda bs=512 count=1 reboot With Debian, FAI, and PXE, this pretty much eliminates us having to touch machines now for anything other than total hardware failure :) I'm not sure such a configuration is possible for the OP, it sounded much like a single system in a co-lo rack somewhere, which makes it difficult. We've 300+ machines of which a random 3-5 seem to be getting re-installed per week. Sorry this probably didn't help much, but I thought some might find the anecdote somewhat interesting. -- Seeya, Paul _______________________________________________ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss
