LILO Re: [Gllug] sync,sync, init 0

[Nix]

[Quotes reordered; please order quotes correctly.]

On Fri, 13 Jul 2001, William Palfreman gibbered:
> On 12 Jul 2001, Nix wrote:
> 
>> sysvinit-2.78 does do its own syncing (see src/halt.c line 237 and
>> src/shutdown.c lines 215 and 263).
>>
>> (However, I still prefer to sync myself in the shutdown scripts; syncing
>> repeatedly does no harm and provides insurance should the behaviour of
>> sysvinit change in the future.)

> Ok then, what happens to, say, a new LILO at this point?  Until I learnt

*confused* LILO has nothing to do with sync-on-termination at all.

> not to do it, I used to do a new LILO, then sync, then reboot, and often
> as not the new LILO  wouldn't get written - disappointing when you wanted
> to play with your shiny new kernel.

The data gets written to the boot= device (see lilo/bsect.c), which is
cached like any other (in the buffer cache, not the page cache); sync()
should write this out as it does anything else.

Your behaviour looks to this innocent like a kernel bug.

>                                      Now I do some mindless disk activity
> before rebooting, like du -h /, and LILO gets written.  What it the Right
> Thing here?

Er, don't do an /sbin/lilo in your shutdown scripts? (I can't imagine
why you'd want to do that, in any case.)

-- 
`I'm not sure whether libtool is an existence proof that you _can_
 write a shell script that handles its arguments correctly, or a
 demonstration that you may try but you are doomed to failure.'
                                                       -- Zack Weinberg


-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug