Changes

Jump to: navigation, search

Talk:Debian template creation

1,297 bytes added, 15:13, 1 December 2013
Path for Lenny now in archive
== edit/symlink /etc/mtab, /proc/mounts ==
Does anybody have valuable information why this would make sense and if this is needed?
http://wiki.openvz.org/Debian_template_creation#Fix_.2Fetc.2Fmtab {{unsigned|Juggler|20:41, 13 February 2007}}
 
: Linking /etc/mtab -> /proc/mounts is needed since VE's root filesystem is actually mounted from the [[Hardware Node]]. Thus, there is no appropriate record in /etc/mtab, so utilities like df fail to work (since they get a list of mounts from the /etc/mtab). After linking /etc/mtab to /proc/mounts, /etc/mtab shows all the mounts, including the one for root filesystem, so df works fine. --[[User:Kir|Kir]] 10:25, 14 February 2007 (EST)
 
== problem with path ==
Hi all,
This is my first trial with openVZ and a wiki. Please give me some hints, if anything is wrong. Don't hesitate to correct my poor english.
I just installed openVZ on a rather old test box with Debian. Installation failed at first and I suppose there is one command missing in a wrong path statement used for the instructions for debianpackages.
When I came to
<code>vzctl start 777</code>
I received the message
<precode>Starting VE ...VE private area /var/lib/vz/private/777 does not exist</precodeMoving the directory 777 from /vz/private to /var/lib/vz/private solves this problem.  Now I could start and enter the new VE. --(unsigned) This isn't a very good place to ask for help. You should use the forum instead. See the link for the forum in the sidebar on the left. --[[User:Dusty|Dusty]] 19:00, 9 April 2007 (EDT) == Debian template creation == For backward compablity to main OpenVZ:
After some researches at the forum I and the wiki I found out that the step Create a symlink from /var/lib/vz to /vz<codepre>vzctl create 777 # ln --ostemplate debian-3.1-i386-minimals /var/lib/vz /vz</codepre> as described in the "OS template cache preparation" is missing in the debian specific section.
Now I could start and enter the new VE== Update document for Debian Etch ==Updated.
One of the first steps I'm not sure how to fit it in the , but if your hardware node is Debian Etch, you don't have to make a temporary VE to create a template. You can just chroot into it instead, which probably saves some hassle. It's probably not worth making a whole new VE should be the installation of security patches via apt-get updatepage for it and it'd probably cause confusion to add it to this document. If someone knows what to do with that information, apt-get upgradefeel free to do something. This fails I always just use a chroot because of missing nameserversit's so much easier for me. But when I added tried it on a RedHat box, it didn't quite work out for some interesting locale reason when installing the line <code>nameserver xxx.xxx.xxx.xxx</code> to /etc/resolv.conf and updates workSSH package.
20
edits

Navigation menu