Host migration from ESXi host to Proxmox host breaks webmin webservice

OS type and version Debian 12
Webmin version 2.111

Hello 2 All,
wanted to migrate a webmin/virtualmin host from a ESXi Server to a Proxmox Server. After that i realized a name change for my network adapter from ens33 to ensp2.
My issue: Now i can not login to the webmin/virtualmin webserver anymore.
Of course i tried to rename the name of the interface in /etc/network/interfaces file but after reboot there is still no access to the webservice possible.
I did not find any documentation how to change the network configuration manually without GUI-Access. What files must be altered to regain access to the webmin/virtualmin GUI?
Greetings - Rudolf

A little bit more information might help us to help you.

@Stegan There was a phonecall during the edit :slight_smile: - you were too fast

1 Like

with the command grep -Rnw ‘/etc/’ -e ‘ens33’ i tried to find all appearances of ens33 - there is also an entry in /etc/webmin/virtual-server/config - but changing this to ensp2 also - it did not work either.

FWIW, when I’ve transferred virtualmin from Vmware to Proxmox, I found it problematic in many ways, especially when using things like lvm or block level storage. After some effort I found the better approach (imo) was to spool up a new vm in proxmox, set it up, install virtualmin using their fantastic script, then backup al the virtual servers from vmware, backup the configuration from vmware, shut that down. On the new proxmox server, sort the IP so it matches, then restore the virtualmin configuration, then restore the servers. Its fairly foolproof and ensures that the underlying linux is at home on KVM with proxmox’s virtual devices. - Just imo.

Fixing your issue on debian 12 would, and this is from memory, ip link set dev ensp2 name ens33 - then restart networking.

1 Like

@GENLTD - Thank you for your shared experience! I will give this a shot!
I will respond the outcome here of course.

Have you tried updating the network interface name in the “/etc/netplan/” directory? That’s where Debian 12 stores its network configuration. You might need to adjust the configuration there to reflect the new interface name. Good luck!

By default Debian 12 doesn’t use Netplan; Ubuntu 18.04 and newer do.

Thank you for your suggestion, but i follow the advice of @GENLTD - but need a time frame for that migration to not disturb the users.
So i will post the outcome of the migration then.
Have a good day!

Before you start the backup of virtual servers, its worth stopping postfix (systemctl stop postfix) which effectively stops email being received whilst you’re backing up (so email’s don’t arrive after you’ve backed up the virtual-server and then be left behind on vmware).

As for timescales, it takes a couple of minutes for each virtual-server to backup to FTP (I do them individually so each virtual server has its own .tar.gz) and about the same to restore each one. The reason I backup each virtual server into its own tarball is because when restoring I do each manually based on ‘importance’ i.e .how much the customer will moan. That way I get the moaners back up first, then work on restoring all the sub domains, alias domains, etc. This keeps downtime really low and the whole process is way faster than trying to migrate the vm.

Do let us know how you get on? it is unfortunate that we’re all now forced to leave vmware which for many years has been a great product.

Thank you for the additional hints @GENLTD !
I am back from a vaccation abroad and after returning i did a new attempt on that topic.
1st step: i did the backup of the virtual servers after closing the external connections to them so that noone is able to alter any data on them.
Meanwhile the Promox PVE got another update from 8.2.1 to version 8.2.2 and my hope was that the migration process would now working better.
2nd step: attach the ESXi datastore to my Proxmox Environment
3rd step: Do the migration task
4th step: Boot the migrated virtualmin server and have a look what the interface name of the network card is. It has been changed to “ens18”
5th step: with nano /etc/network/interfaces => there was “ens33” configured and i changed all the ens33 entries to “ens18”
6th step: i restatred the network with systemctl restart NetworkManager.service and check the change
7th step rebooted the virtualmin server and enjoyed a fully working migrated virtualmin server!

Thank you all about caring about the topic!
@GENLTD : Leaving the Vmware/ESXi environment is unfortunate to us because we are so used to work with it. But which each day i am getting more and more exited with the proxmox environment and their community so i am looking forward to this new prosperous relationship.

How does the performance compare?

@Randomz : i did no extensive performance comparisons. The working functionality is momentarily in the focus. The overall performance for this type of vm, a webserver, is sufficient to me. When it comes to performance matters - there are so many aspects to look at. Also what type of performance is in the focus. Different needs need different approaches. It maybe starts with the chosen filesystem. You have so many options in proxmox to vary the datasets on a proxmox node that you can have different partitions with different filesystem performance options and so forth.
My summary is: Proxmox have mighty possibilities but it has to be my task to sharpen my capabilities to fruitfully use them.

This topic was automatically closed 8 days after the last reply. New replies are no longer allowed.