Re: automatic reboots

Sean Utt (sean@northwest.com)
Sun, 29 Dec 1996 19:26:11 -0800 (PST)


On Sun, 29 Dec 1996 tytso@mit.edu wrote:

> Date: Fri, 27 Dec 1996 09:53:34 -0800 (PST)
> From: Sean Utt <sean@northwest.com>
>
> add a -y to the line in your /etc/rc.d/rc.S
>
> Using fsck -y will work, but e2fsck is designed to stop when there's a
> chance that a system administrator can do a better job than simply just

I am aware of that, and I should have warned as you did. The servers I am
working with are located where direct access is not convenient, and there
have been times where a lockup would require a remote power down to bring
the server back up. In this case, the server stopping to wait for me to
drive 15 miles to run fsck is not acceptable. The system being on a UPS
(which it is) is not a factor in this case. So, if you really care about
your filesystems, have a UPS, and run regular backups :-)

On a related topic....

Is there any possibility that a true serial boot at the "lilo" level is
going to be available any time soon for linux on intel platforms? That
would make this less of a problem for all us remote administrators. The
ability to dial in and boot into single user mode would be helpful in
itself, but the option of booting different kernels remotely, and
interracting with the system during boot would be even more helpful.

For linux to make inroads against the other unix platforms commercially it
seems like a necessity.

sean

> answering yes to all of the questions. If you do this, I hope it is on
> a server machine where you can easily duplicate all of the filesystems
> if something goes wrong and fsck -y just makes things right.
>
> If you really care about contents of your filesystem, then get a UPS.
> If weight is more important than reliability, then perhaps you can skip
> the UPS.
>
> - Ted
>
>