disk inconsistency - should 'fsck' always executed with -y ?

[glenn at aoi.ultranet.com (Glenn Burkhardt)]

> Hmmm....  You would need to specify all of your filesystems on the command
> line... Generally if the filesystem IS modified, you should reboot
> afterward, in order to avoid buffer cache from corrupting the fixed
> filesystem and such.  I'm not positive how your system will behave if you
> do this since I've never messed with it, but I think if you modify your
> start scripts to do what you suggest, they will simply plow on through and
> bring the system up after fixing problems, which could result in you
> running on a corrupted filesystem.

Good point.  However, e2fsck has specific return codes for "system should
be rebooted", which I presume the startup scripts already check for, and
handle appropriately (or could be made to do so).

Thanks for the book reference - I'll check it out.  The Linux SAG doesn't
say very much.  I'll probably also pull the fsck code and see what I can find.

Philosophically, I'm still of the mind that the system ought to just fix 
itself.  I knew a student when I was in grad school that would correct 
file system problems manually on the PDP-11 Unix, but never really figured 
out what he was doing. 

Probably, there are few if any people who really know how to fix a filesystem
"manually", and tell 'fsck' not to fix problems.  This is probably just
another hangover from the early Unix days.

-
Subcription/unsubscription/info requests: send e-mail with
"subscribe", "unsubscribe", or "info" on the first line of the
message body to discuss-request at blu.org (Subject line is ignored).