RE: locating the source for hdd activity
Did you try the find command I suggested?
Try something like (in a script, run as root):
sleep 310; find / -mount -type f -mmin -5
find /usr -mount -type f -mmin -5
find /home -mount -type f -mmin -5
Run it, then leave your computer alone until the script complete. It should
list all files modified in the last 5 minutes. I was able to figure out
samba was active every couple of minutes that way. Most likely you'll find
something in /var is getting touched.
It is unlikely that something is reading a file over and over but it might
be possible. You could temporarily reenable atimes and change the find
command to see if it finds anything.
There is also the possibilities that something is just creating and deleting
a file in which case you could remove the '-type f' from the find commands
above to hunt for directories as well.
And there is always 'lsof'. That might help you find something.
> -----Original Message-----
> From: Jeff [mailto:jcoppock1@attbi.com]
> Sent: Monday, November 18, 2002 5:22 PM
> To: debian users
> Subject: Re: locating the source for hdd activity
>
>
> Charlie Reiman, 2002-Nov-18 11:19 -0800:
> > I've got my hd powering down just fine on my laptop with ext3.
> Use noatime,
> > as noted. You will also need to disable syslog marking (see
> > /etc/init.d/sysklogd, add "-m 0" as an option) and enable noflushd
> > (/etc/default/noflushd). This is with debian sid or woody,
> 2.4.18 kernel.
> >
> > Another tip: once you are mounting noatime you can use find to
> hunt down any
> > files being modified. Leave you laptop on but untouched for 20
> minutes or
> > so, then try (assuming only one partition):
> >
> > find / -mount -type f -mmin -5
> >
> > Will find everything modifed in the last five minutes.
>
> This thread encouraged me to do this on my laptop, but something is
> still causing it to not work.
>
> - I set noatime for all my ext3 partitions and this shows when doing
> the "mount" command.
>
> the pertinent devices:
> /dev/hda6 on / type ext3 (rw,noatime,errors-remount-ro)
> /dev/hda4 on /usr type ext3 (rw,noatime)
> /dev/hda7 on /home type ext3 (rw,noatime)
>
> - I set the "-m 0" for sysklogd and this shows in the process view
>
> 190 /sbin/sysklogd -m 0
>
> - when i run noflushd -d -t 1 /dev/hda
> it spins down after 1 minute, but 30 seconds later it spins back up.
>
> How can I find out what is causing this?
>
> thanks,
> jc
>
>
> --
> Jeff Coppock Systems Engineer
> Diggin' Debian Admin and User
>
Reply to: