VE 5 installation fails right at the end

[philled]

I'm installing VE 5 on a machine with 2 hard disks and choosing ZFS0 as the file system. I erased all partitions off the disks before the installation. The installation seems to run smoothly right until the very end when I get this error which I can see when I press Ctrl-Alt-F2:

command 'chroot /rpool/ROOT/pve-1 dpkg --force-confold --configure -a' failed with exit code 1 at /usr/bin/proxinstall line 385
umount: /rpool/ROOT/pve-1/var/lib/vz: not mounted

I've tried multiple times. Can anyone advise how to fix this? I used the same installation CD to install onto a different machine a few weeks ago and it all worked fine.

 

[philled]

I've been testing some more on this. I cannot get the installation to complete with ZFS0, but if I choose EXT4 it completes successfully. But why? What's the problem with ZFS here? I'd like the installation to be ZFS0 not EXT4.

 

[dietmar]

But why? What's the problem with ZFS here? I'd like the installation to be ZFS0 not EXT4.

Please inspect the install log in /tmp/install.log - that should contain all details.

 

[philled]

Hi, I looked at /tmp/install.log but it's not really helping. I've copied a picture of the end of the log file to http://picpaste.com/Proxmox_Install_log-JaPHYKsb.jpg

Does this shed any light on why the installation is failing?

 

[dietmar]

Does this shed any light on why the installation is failing?

No - the error must be earlier ...

 

[philled]

No - the error must be earlier ...

Hi, I've got the install.log file off the machine. It's too large to paste here so I've put it up at http://www.oaklands.aust.com/Proxmox_install.txt.
Can you see anything there that looks suspicious? At line 17082 this happens which doesn't look good:

Setting up libfaketime:amd64 (0.9.6-7+b1) ...
Setting up libldap-2.4-2:amd64 (2.4.44+dfsg-5) ...
Setting up python3-chardet (2.3.0-2) ...
Segmentation fault
dpkg: error processing package python3-chardet (--configure):
subprocess installed post-installation script returned error exit status 139
Setting up novnc-pve (0.6-4) ...
Setting up libasprintf0v5:amd64 (0.19.8.1-2) ...
Setting up liburi-perl (1.71-1) ...

One thing I wondered was whether it could be a partitioning issue of some sort - previously the 2nd disk was a LVM partition and I wonder if there's a problem now making it a ZFS partition. Before the install I used Gparted to remove all the partitions but it didn't help.

Or some sort of problem updating /boot or grub perhaps?

 

[dietmar]

Seem that "Segmentation fault" is the problem. But I have never seen that before, and I have install pve many times using zfs.
What kind of hardware do you run on (prozessor/board)?

 

[philled]

What kind of hardware do you run on (prozessor/board)?

It's an HP XW workstation type PC with a couple of disks, one Seagate and can't remember the other, one on board NIC and an HP dual NIC as well.

What I ended up doing is installing v4.4 instead. That worked and then I used apt to upgrade to v5.0. Odd that 4.4 worked but not 5.0.