Subject: Re: fsck issue
To: Michael L. VanLoon -- HeadCandy.com <michaelv@HeadCandy.com>
From: Markus Illenseer <markus@server.peacock.de>
List: current-users
Date: 08/06/1996 07:35:47
> >If you're running with securelevel > 0, you won't be able to write to
> >a mounted partition. If not, you could `mount -u -r' it to downgrade
> >the mount to read-only while you run fsck(8), and then do a `mount -u
> >-o reload' when you're done. (I'm not actually sure whether the `-o
> >reload' is necessary there, but it wouldn't hurt.)
>
> Of course, the problem with that being that a process that tries to
> write to that filesystem during that period of time will get an error,
> instead of just blocking until the check is done, which would be the
> desired behavior.
Yes, you are right. For most purposes the suggestion from Charles seems
to be sufficient though. I must try it out.
Even though Unix is more than 2 decades old, it seems there is still
no feasible mechanism to do a system check w/o being forced to shutdown
to single user mode or umount all partitions. umounting for me means
that I cannot continue to work via xterm, and have to go downstairs
to the console... no, i am not lazy :-)
--
Markus Illenseer