Problem with pve-upgrade-1.9-to-2.0

[stef1777]

Hello!

I tried the upgrade script on a testing 1.9 node.

The script failed because the script configure /etc/apt/sources.list with non available repository.

W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/main/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]
W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/contrib/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]
W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/non-free/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]

I manually removed the volatile line and the script restarted fine.

And stopped later on this:

Processing triggers for initramfs-tools ...
update-initramfs: Generating /boot/initrd.img-2.6.32-7-pve
W: mdadm: /etc/mdadm/mdadm.conf defines no arrays.
W: mdadm: no arrays defined in configuration file.
Errors were encountered while processing:
procps
hdparm
E: Sub-process /usr/bin/dpkg returned an error code (1)
minimal upgarde failed

Seems the script answer yes to replace config file by default and broke the procps and hdparm config.

Manually run dpkg --configure solve the problem.

And again stop:

Errors were encountered while processing:
cron
udev
keyboard-configuration
console-setup
ntp
at
nfs-common
openssh-server
bind9
postfix
apache2.2-common
apache2-mpm-prefork
apache2
libapache2-mod-perl2
libapache2-mod-apreq2
libapache2-request-perl
E: Sub-process /usr/bin/dpkg returned an error code (1)
dist-upgrade failed

Upgrade from 1.9 seems very hot...




pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.32-6-pve
proxmox-ve-2.6.32: 1.9-55+ovzfix-2
pve-kernel-2.6.32-4-pve: 2.6.32-33
pve-kernel-2.6.32-6-pve: 2.6.32-55+ovzfix-1
pve-kernel-2.6.32-7-pve: 2.6.32-55+ovzfix-2
qemu-server: 1.1-32
pve-firmware: 1.0-15
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: 3.0.29-3pve1
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.15.0-2
ksm-control-daemon: 1.0-6

 

[tom]

Hello!

I tried the upgrade script on a testing 1.9 node.

The script failed because the script configure /etc/apt/sources.list with non available repository.

W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/main/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]
W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/contrib/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]
W: Failed to fetch http://volatile.debian.org/debian-volatile/dists/squeeze/volatile/non-free/binary-amd64/Packages 404 Not Found [IP: 2001:858:2:2::2 80]

no, volatile was configure by you or someone else, Proxmox VE never used this. looks like you got a customized installation. Remove all 1.9 package manually with apt, do a dist upgrade and install 2.0. or do a re-install.

pve-manager: 1.9-26 (pve-manager/1.9/6567)
running kernel: 2.6.32-6-pve
proxmox-ve-2.6.32: 1.9-55+ovzfix-2
pve-kernel-2.6.32-4-pve: 2.6.32-33
pve-kernel-2.6.32-6-pve: 2.6.32-55+ovzfix-1
pve-kernel-2.6.32-7-pve: 2.6.32-55+ovzfix-2
qemu-server: 1.1-32
pve-firmware: 1.0-15
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: 3.0.29-3pve1
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.15.0-2
ksm-control-daemon: 1.0-6

just an additional note, your 1.9 is not up2date. a requirement for the upgrade script is that you run latest 1.9, see the upgrade docs - http://pve.proxmox.com/wiki/Upgrade_from_1.9_to_2.0#Requirements

 

[dietmar]

Seems you have not read the upgrade instructions at: http://pve.proxmox.com/wiki/Upgrade_from_1.9_to_2.0

"Upgrades should only be done on non-modified systems"!

 

[stef1777]

You're right. apt.source was modified by hand. I removed the added lines. The rest is standard but the upgrade is completely unstable now.

Errors were encountered while processing:
pve-cluster
redhat-cluster-pve
fence-agents-pve
libpve-access-control
clvm
libpve-storage-perl
qemu-server
resource-agents-pve
pve-manager
vzctl
proxmox-ve-2.6.32
E: Sub-process /usr/bin/dpkg returned an error code (1)
minimal upgarde failed

1.9 was updated to be sure to have the last 1.9 release just before running the upgrade script.

 

[dietmar]

You need to resolve those conflicts manually

 

[stef1777]

Yes, but I don't find the right way. I tried dpkg --configure of each package manually and each time I've a dependency problem. Bad loop!

Sure that in the real life, I will prefer to reinstall from iso and reload saved VM.

 

[dietmar]

Yuu can simply try to remove conflicting packages. The upgrade script will install them later anyways.

 

[stef1777]

I've found the blocking package. Do you know why "Unable to get local IP address"

Setting up pve-cluster (1.0-26) ...
Starting pve cluster filesystem : pve-cluster[main] crit: Unable to get local IP address
(warning).
invoke-rc.d: initscript pve-cluster, action "start" failed.
dpkg: error processing pve-cluster (--configure):
subprocess installed post-installation script returned error exit status 255
Errors were encountered while processing:
pve-cluster

 

[dietmar]

I've found the blocking package. Do you know why "Unable to get local IP address"

You should have an entry in /etc/hosts (for the name used in /etc/hostname).

Note: if you post both files if can tell you how to fix that.

 

[stef1777]

Oops, seems that something emptied /etc/hosts. The file was completely empty. I corrected that.

After, that upgrade well finished.

 

[stef1777]

Hello again me!

I finally well upgraded 2 testing nodes with success. Apart some problems during the apt-get dist-upgrade, the rest is fine.

I've a new problem.

sd-33092:~# ./pve-upgrade-1.9-to-2.0 --import
starting import
import file: /etc/pve/user.cfg
unable to open file '/etc/pve/user.cfg' - Permission denied at ./pve-upgrade-1.9-to-2.0 line 125.

The node contains one openvz container. The file /etc/pve/user.cfg don't exist. /etc/pve.old-1.9/ exist.

I don't know how to alive the container.

 

[dietmar]

Do you run a cluster, or a single node?

 

[stef1777]

Single.

The network here, doesn't support multicast.

 

[dietmar]

What is the output of

# ls -l /etc/pve

 

[stef1777]

Here it goes!


sd-33092:~# ls -l /etc/pve
total 1
-r--r----- 1 root www-data 291 Apr 3 14:59 cluster.conf
lr-xr-x--- 1 root www-data 0 Jan 1 1970 local -> nodes/sd-33092
lr-xr-x--- 1 root www-data 0 Jan 1 1970 openvz -> nodes/sd-33092/openvz
lr-xr-x--- 1 root www-data 0 Jan 1 1970 qemu-server -> nodes/sd-33092/qemu-server

sd-33092:~# ls -l /etc/pve.old-1.9/
total 40
-rw-r--r-- 1 root root 1457 Nov 14 13:00 cluster.cfg
drwxr-xr-x 2 root root 4096 Apr 3 14:24 master
-rw-r--r-- 1 root root 26 Nov 14 11:25 pve.cfg
-rw------- 1 root root 887 Nov 14 10:39 pve-root-ca.key
-rw-r--r-- 1 root root 1180 Nov 14 10:39 pve-root-ca.pem
-rw------- 1 root root 3 Nov 14 10:39 pve-root-ca.srl
-rw------- 1 root root 887 Nov 14 10:39 pve-ssl.key
-rw-r--r-- 1 root root 956 Nov 14 10:39 pve-ssl.pem
-rw-r--r-- 1 root root 13 Nov 14 11:25 qemu-server.cfg
-rw-r--r-- 1 root root 134 Nov 14 13:03 storage.cfg

Just to know: On the 2 nodes upgraded, I have to reinstall grub. The server was no longer booting after the upgrade.

 

[dietmar]

Please run

# pvecm updatecerts

and try again

 

[stef1777]

sd-33092:~# pvecm updatecerts
no quorum - unable to update files

sd-33092:~# ./pve-upgrade-1.9-to-2.0 --import
starting import
import file: /etc/pve/user.cfg
unable to open file '/etc/pve/user.cfg' - Permission denied at ./pve-upgrade-1.9-to-2.0 line 125.

No luke!

 

[dietmar]

That still looks like the hostname bug. What is the output of:

# /etc/init.d/pve-cluster stop
# /etc/init.d/pve-cluster start

 

[stef1777]

No output message.

/etc/hosts and /etc/hostname are right. But when I've upgraded the server, hosts was empty. I corrected that during the upgrade.

 

[dietmar]

You claim you have a single node, but I just saw that you have a cluster configuration file '/etc/pve/cluster.conf'?!