Unofficial updater script for Proxmox VE (HOST / LXC / VM )

[BassT]

v4.3 released now

https://github.com/BassT23/Proxmox

 

[a13xde]

Hi @BassT,

First: thanks for sharing your Ultimate Updater Script !

I want to update my LXC containers with it (without the hosts), I have a 3 node cluster and on 2 of the nodes the script runs through the Containers and does what it should, but when it tries to connect to the 3rd node I always get the error message:

Host key verification failed.
scp: Connection closed
Error during Update --- Exit Code: 255

If I try to ssh from the host with the updater Script to node3 it works, also if i try to manually send a file from node1 to node3 with hostname and/or IP address I don't get this "Host key verification failed" Error.

I don't know what the script does differently here than when I run ssh or scp manually, do you have any ideas?

 

[BassT]

Hi @a13xde,

thanks for using my script

You need better read the info here The last part is the one you are serching for.
The script run from node1 -> node2 and after that from node2 -> node3

So you must check the ssh connection from node2 to node3

 

[a13xde]

Thanks for the hint @BassT next time I really need to read more carefully
Nevertheless, an ssh and scp from node2 -> node3 also works and does not produce the error message I receive from the script. Can you tell me the exact command I have to test ? I tried "ssh <IP-of-node3>" and "scp <Testfile> <IP-of-node3>:/root/<Testfile>"

/etc/hosts are the same on all nodes and checked the IPs with

awk '/ring0_addr/{print $2}' "/etc/corosync/corosync.conf"

 

[BassT]

@a13xde

with this command you can list the used "hopping"

awk '/ring0_addr/{print $2}' "/etc/corosync/corosync.conf"

Here shout come out:

<node1 ip>
<node2 ip>
<node3 ip>

You can test if the connection work from node2 with

ssh <node3 ip> uptime

for example

 

[a13xde]

@a13xde
You can test if the connection work from node2 with

ssh <node3 ip> uptime

for example

Yeah, this works from each node to each of the the other nodes with hostname and IP adresses listed in /etc/corosync/corosync.conf as ring0_addr.
I had run:

pvecm updatecerts -f
systemctl restart pvedaemon pveproxy

on each node and doublechecked ssh login again , without the error.

i have no idea anymore. I'm still receiving the error when the script tries to connect from node2 to node3:

--- Finished extra updates ---

--- Check Status for Welcome-Screen ---
--- Finished check ---


[Info] Skipped all VMs by the user

Host key verification failed.
scp: Connection closed
Error during Update --- Exit Code: 255

There are an error during update
Please check Output here or Logfile

no more Log in the Logfile

 

[a13xde]

For all those interested in the solution:
my second node in the 3 node cluster did not know his own rsa key and this was the problem for me.

"ssh node2hostname uptime" from node2 with adding key to known hosts resoved the issue for me.

So if you receive an "Host key verification failed" error, it's best to check not only ssh to the other nodes, but also to itself

Many thanks to @BassT helping me to troubleshoot this

 

[louie1961]

I think I would much prefer Ansible for doing this. It affords much more control.