[Quantian-general] 7.9.1 and 2.4.27 kernel
Dirk Eddelbuettel
edd at debian.org
Thu Jan 5 02:31:06 UTC 2006
On 4 January 2006 at 09:16, Christopher Heiny wrote:
| [Sorry Dirk - kmail is doing something really odd with addressing and
| identities on messages from quantian-devel. I'm resending this one.]
|
| On Wednesday 04 January 2006 07:15 am, picked up the following
| transmission from Marco Caliari:
| > Here are my results about 2.4.27-om kernel and unionfs.
| > I remember you that I'm using "boot Quantian directly from hard disk
| > via the grub bootloader".
| >
| > 1) When the Quantian iso is on a Ext3 filesystem, no problems at all:
| > openMosix starts and the two KNOPPIX* images are mounted (both with
| > Dirk's unionfs.o and version 1.0.14).
| >
| > 2) When the Quantian iso in on a ReiserFS filesystem, the two images
| > are mounted, but openMosix does not start. However, after a
| >
| > /etc/init.d/openmosix restart
| >
| > openMosix is up. From a df
| >
| > root at 0[init.d]# df
| > Filesystem 1K-blocks Used Available Use% Mounted on
| > /dev/root 9677 30 9147 1% /
| > /cdrom2.loop/Q*.iso 2506788 2506788 0 100% /cdrom/
| > /dev/cloop 5678206 5678206 0 100% /KNOPPIX
| > /ramdisk 1633392 5556 1627836 1% /ramdisk
| > /UNIONFS 8382300 6754464 1627836 81% /UNIONFS
| > /cdrom2.loop/Q*.iso 2506788 2506788 0 100% /cdrom/
| > /dev/cloop 5678206 5678206 0 100% /KNOPPIX
| > /dev/cloop2 1070706 1070706 0 100% /KNOPPIX2
| > mfs_mnt 9999999 -18446744073619551616 9999999 101%
| > /UNIONFS/mfs
| >
| > the last row seems strange to me.
|
| That whole df output looks kind of odd to me - several mount points are
| listed twice. But maybe that's normal when booting from a hard drive
| or maybe I just don't understand UNIONFS.
It doesn't look that way under kernel 2.6 either when unionfs works normally.
The equivalent would be something like this (judging from what I see under
vmplayer; I'd have to reboot otherwise)
root at 0[init.d]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/root 9677 30 9147 1% /
/dev/hdc 2506788 2506788 0 100% /cdrom/
/ramdisk 1633392 5556 1627836 1% /ramdisk
/UNIONFS 8382300 6754464 1627836 81% /UNIONFS
/dev/cloop 5678206 5678206 0 100% /KNOPPIX
/dev/cloop2 1070706 1070706 0 100% /KNOPPIX2
So so gar Marco is still the only one who claims that 2.4, oM and unionfs can
play along. We'd really need two (or more) machines in a small cluster to
share processes to be sure.
| I didn't get a chance to run any tests last night - maybe tonight. I'll
| check the df output then, as well.
Sounds good.
Regards, Dirk
--
Hell, there are no rules here - we're trying to accomplish something.
-- Thomas A. Edison
More information about the Quantian-general
mailing list