On Sun, 2005-04-24 at 22:08 +0100, Nigel Kendrick wrote:
> Here's one for you if you have the time:
>
> I very stupidly killed the wrong process on a server on a remote site on
> Friday and now I cannot ssh login to it! It's not a major problem as it's
> main functions are as a file and print sharer (samba) and to run
> postfix/mailscanner, and I can have someone on site reboot the server when
> they start work on Monday, but I was wondering whether I could 'get in' by
> other means to reboot it - so far I have tried the following without luck:
>
> Simultaneous login requests to 'reach' a listening login as I think I've
> only screwed up pts0
>
> A VNC connection to a machine on site and then using telnet (nope, telnet
> disabled)
>
> A VNC connection to a machine on site and then a VNC connection to the
> server (VNC not running on this one).
>
> Any suggestions to issue a remote reboot - as I said, it's not a major issue
> as I can have the server rebooted by someone on site.
>
> Thanks
>
> Nigel Kendrick
>
Serial connection are most useful in these cases, but after the fact
this is not very helpful. I have also used cron jobs when doing remote
admin on machine with good success, if changing firewall configs, I
would revert he change and restart iptables, doing ssh configs same
deal.
Another good thing is to have a generic local account that has sudo
privileges assuming you have someone that can hope of the console for
you.
As long as the changes you made did not toast anything a reboot is your
likely only recourse.
Ted
