Solved: Linux dies at 6am....

[Damien Elmes]

Damien Elmes <resolve at repose.cx> writes:

> James Macnicol <jamesm at ee.adfa.edu.au> writes:
> 
> > 	Thanks to everyone who replied, especially those who pointed
> > out that cron.daily is run around the time I noticed the machine died.
> > This was exactly what I needed.  I had a look at the things that
> > cron.daily did and noticed that some of them took a fair amount of
> > time, my guess was that there were some things there that were
> > traversing the entire filesystem; probably all filesystems.  It then
> > occured to me that I was mounting my W2K partition (read-only) and the
> > Linux NTFS driver is a piece of cr at p and that the reason the system
> > was locking up was a cron job that was going through the mouted
> > filesystem was enough to cause the NTFS driver to lock up the
> > kernel....  So I unmounted that filesystem before I left the machine
> > last night and it's still up this morning.  This isn't an acceptable
> > solution given I need access to the data on there but I should be able
> > to figure out pretty easily how to stop whatever cron job it is from
> > touching that partition.
> 
> /etc/updatedb.conf
> 
> add the path to this line:
> 
> PRUNEPATHS="/tmp /usr/tmp /var/tmp /afs /amd /alex /var/spool"

(that or you can add ntfs to the list of filesystem types to prune)

> cheers,

-- 
Damien Elmes
resolve at repose.cx