On Sunday 02 October 2005 21:52, nordi wrote:
> Those a pretty big differences, 63.7 vs 68.1 seconds seems like
> something is making the measurements rather unprecise. My measurements
> were 60.06, 58.26, 60.24 and 59.21. So the differences were _much_ smaller.
>
> Could it be that some strange hardware or a slow dhcp server is the
> reason for this deviation? I measured the whole thing with network cable
> unplugged and no extra hardware attached, trying to eliminate any
> hardware-dependant effects.
Yes, definitely network issues. Thus I deactivated all network interfaces,
unplugged the machine, and rebooted a couple of times.
without modification
58.48 33.87
58.52 33.25
60.92 33.17
average: 59.3
with modification
59.00 32.99
59.94 32.72
58.05 32.21
average: 59.0
This time there are no big differences anymore but the added preload does seem
to have nearly no effect. To unmake the hack I just did
--- boot.localfs.modified 2005-10-01 20:09:54.000000000 +0200
+++ boot.localfs 2005-10-03 11:36:46.000000000 +0200
@@ -138,7 +138,6 @@
mount -fv -t devpts $optpts devpts /dev/pts
rc_status
mount -av -t nonfs,noproc,nodevpts,nosmbfs -O no_netdev
- /usr/bin/ionice -n6 /sbin/preload < /var/cache/preload/kde.preload &
rc_status -v1 -r
;;
stop)
Cheers,
Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
