INCORRECT BLOCK COUNT and UNALLOCATED errors

[Lou Hevly]

At 09:52 03/11/06 -0500, Woodchuck wrote:
>On Thu, 2 Nov 2006, Lou Hevly wrote:
>
> > But then I restarted Apache and I got errors again.
>
>I don't think you can run fsck on a mounted filesystem and ~not~
>get errors.  chrooted apache will doubtless have activity on /var.
>(Also syslogd will, too, at least intermittently).
>
>I wouldn't pay attention to errors found by fsck on a mounted system.

Indeed.  I did finally figure this out after an embarrassingly long 
while.  Luckily this is a newbies list.

I'd like to say that it's certainly a tribute to OpenBSD (and djb) that 
someone who knows as alarmingly little about computer fundamentals as I 
do can have kept a web-server running since 2.7 with no significant 
problems.


> > >However,  you will need to isolate the system and perhaps kill 
> various
> > >processes. (Those with files open on /var or whatever filesystem 
> you
> > >are fsck'ing).   Do not be umounting /var or /home on a system 
> with
> > >random users and processes.  You want a quiet system.
> >
> > My system is pretty calm.
>
>You want it completely calm ;-)

;-)


>Use fstat to see if anything is opened on a filesystem, then umount
>it before fsck'ing.
>
>SYSV style init had "runstates", one of the ones BSD doesn't have
>kicked off all users except root from interactive sessions and killed
>off daemons.  But I believe this state also closed the network.  You'd
>like the system to be in a state in which only root can have an 
>interactive
>process, all daemons were killed, but in which enough networking was
>up so that you could allow the ssh process of root to endure.

Thanks for your comments.  If you're ever in Barcelona let me know and 
I'll buy you a brewski.