[SOLVED] Changed VMBR0 IP Address Before Setting Up Cluster and Experienced Issues

[mhayhurst]

Hello everyone,

I corrected this issue but wanted to share my experience with setting up a cluster following the Proxmox Wiki after I had already changed the IP address of vmbr0 on both Proxmox hosts. I installed Proxmox on two platforms with hostnames: proxmox1 and proxmox2. Both of the machine's IP addresses were set during the installation: proxmox1: 192.168.1.2 and proxmox2: 192.168.1.3. I set up a cluster using these two machines last night but a few weeks prior I had changed the IP addresses of proxmox1 to: 192.168.1.4 and proxmox2 to: 192.168.1.5 as well as performed a reboot on both machines. I did this through the Proxmox webgui by editing the vmbr0 IP address on both machines. I thought this would be all that was needed but apparently changing the vmbr0 in the webgui does not change the: /etc/hosts IP address. So when I went to setup the cluster I ran into errors with the corosync service. I also found that the: bindnetaddr in /etc/corosync/corosync.conf and /etc/pve/corosync.conf were both set to proxmox1's previous IP address (192.168.1.2). Also, changing the bindnetaddr in /etc/corosync/corosync.conf on both machines would not suffice as it reverts back to 192.168.1.2 after a reboot or service restart of corosync...even though I had changed the /etc/hosts on both machines to their new IP address. In the end I had to do the following on both machines:

1. Change /etc/hosts on each machine to reflect their new IP address
2. Executed:

   service corosync stop

   systemctl stop pve-cluster

   pmxcfs -l

3. Edited the bindnetaddr in /etc/corosync/corosync.conf & /etc/pve/corosync.conf on both machines to reflect the new IP address of: proxmox1 (192.168.1.4).
4. I then rebooted both machines but was getting a: "Connection Error" in the webgui on proxmox2
5. I checked in /var/log/syslog on proxmox2 and saw this error:
   

   proxmox2 pveproxy[3120]: /etc/pve/local/pve-ssl.key: failed to load local private key (key_file or key) at /usr/share/perl5/PVE/HTTPServer.pm line 1626

6. I was able to remedy this error by executing:

   pvecm updatecerts

   and then I performed a reboot on proxmox2.

So here are my questions:

1. Should changing the vmbr0 IP address in the webgui also update the /etc/hosts file?
2. Why are there two corosync.conf files in two different locations?
3. Is there anything else I should do on each machine as good measure considering the changes I made?

 

[manu]

1. no, you have to update /etc/hosts manually
you the aim of /etc/hosts it to provide local name resolution
2. /etc/corsync/corosync.conf is used at boot and kept it sync afterwards with /etc/pve/corosync.conf
if you need to edit this file, have a look at http://pve.proxmox.com/pve-docs/chapter-pvecm.html#_corosync_configuration
3. if the output of
pvecm status
service pve-cluster status

are both OK you should be good with your changes

 

[mhayhurst]

1. no, you have to update /etc/hosts manually
you the aim of /etc/hosts it to provide local name resolution
2. /etc/corsync/corosync.conf is used at boot and kept it sync afterwards with /etc/pve/corosync.conf
if you need to edit this file, have a look at http://pve.proxmox.com/pve-docs/chapter-pvecm.html#_corosync_configuration
3. if the output of
pvecm status
service pve-cluster status

are both OK you should be good with your changes

I thought I replied to this earlier but this fixed the problem, thank you!