HOWTO: chroot - admin [Ubuntu] Linux on another partition

[charlieg]

I have two linux partitions - a handy set up that allows me to admin one from the other should something go wrong (e.g. with an upgrade) or I want to install something without booting into the other. I have discovered this through various tutorials / docs scattered around the web but nothing had them all in one place so here it is: charlieg's ace and simple guide to using chroot

This is a very handy way to update between major distro versions and I have done it for Gutsy->Hardy - if it doesn't boot, you can boot into the working partition and solve the problem from there!
The following howto should apply to most Linux distributions. You should even be able to boot into a liveCD and do this from the console.
You will need to install 'chroot', and 'mount' should be there by default.
You gotta be root (sudo is ok)

So, here it is:
[charles@localhost ~]# sudo -s -H
Password:
[root@localhost ~]# mount --bind /dev/ /media/ubuntu/dev
[root@localhost ~]# mount --bind /dev/pts /media/ubuntu/dev/pts
[root@localhost ~]# mount --bind /dev/shm /media/ubuntu/dev/shm
[root@localhost ~]# chroot /media/ubuntu
root@localhost:/# mount -t sysfs sysfs /sys
root@localhost:/# mount -t proc proc /proc
And that's it! You're in! The environment is set up to basically use the core devices from your active partition. Now you should be able to admin-away without trouble.
root@localhost:/# aptitude update && aptitude dist-upgrade

[charlieg]

Oops, the 'su -' command is what I use on my Fedora partition - from Ubuntu it has to be 'sudo -s -H' to get a root capable terminal. Corrected.

[az]

I have always done it simpler:

If the chroot directory is called chroot:

sudo chroot chroot
mount /proc
mount /sys
mount -t devpts none /dev/pts

and then do whatever you want...


To leave, you need to run:

umount -lf /proc
umount -lf /sys
umount -lf /dev/pts
exit

[alSee]

Console applications work fine by this method. But GUI ones generate such error:
xpsgui: cannot connect to X server :0.0
I use the temporary solution of starting vnc-server in chroot and vnc-client in current environment. But it isn't true way.
Any suggestions?

[octoberdan]

Thank you for the tutorial! I was running Slackware based environment (Backtrack3) chrooting into an Ubuntu environment. However, I ran into some difficulties regarding network connectivity; the chroot environment didn't have an IP address but couldn't get one as the Slackware environment had a dhcp client running. Here's what happened:


/# dhclient
Internet Systems Consortium DHCP Client V3.1.1
Copyright 2004-2008 Internet Systems Consortium.
All rights reserved.
For info, please visit http://www.isc.org/sw/dhcp/

Listening on LPF/eth1/00:08:74:e5:90:6b
Sending on LPF/eth1/00:08:74:e5:90:6b
Listening on LPF/eth0/00:16:6f:55:68:77
Sending on LPF/eth0/00:16:6f:55:68:77
Can't bind to dhcp address: Address already in use
Please make sure there is no other dhcp server
running and that there's no entry for dhcp or
bootp in /etc/inetd.conf. Also make sure you
are not running HP JetAdmin software, which
includes a bootp server.


Oh nose! So I went into the host machine (non-chroot environment) and found the process of the dhcp program and killed it:


bt ~ # ps aux | grep dhcp
root 7830 0.0 0.0 1676 224 ? Ss 22:18 0:00 /sbin/dhcpcd -t 60 eth0
root 8413 0.0 0.2 1760 512 pts/2 R+ 22:54 0:00 grep dhcp
bt ~ # kill 7830


In other environments, the dhcp program may be different. Debian based operating systems (such as Ubuntu) usually use dhclient.

Hope this info helps someone...

[justin.perkins]

thanx for this, cheers. :) using it to mount a usb-fs which I had loaded up into ramdisk, to update/switch contexts while running from ram, to disk, then exit back into the ramdisk version. handy. :)

[nothingspecial]

old thread.

Closed.