Upgrading a Proxmox 8.1.3 cluster to version 8.2

[punkn123]

Hello There,

I'm new to Proxmox and currently managing a Proxmox cluster with four nodes, each hosting important VMs. Is there a way to upgrade the nodes one by one without disrupting the cluster? Additionally, after upgrading a node, will I need to reboot it? Rebooting isn't very convenient for me.

I'm planning to use the following commands for the upgrade:

apt-get dist-upgrade

or

apt full-upgrade

.

Any advice would be greatly appreciated!

Thanks!

 

[Falk R.]

I'm planning to use the following commands for the upgrade:

apt-get dist-upgrade

or

apt full-upgrade

.

Both does the same, but the second one ist the newer command. You can also upgrade via GUI.

In a cluster, the nodes should always be updated one after the other. This always works without interruption. Kernel updates should always be restarted, otherwise they cannot be applied.

In a real cluster, shared storage or ZFS replication is used, so the VMs can be migrated during operation. This also happens automatically with activated and correctly configured HA.

VMs can also be migrated without shared storage, but then the disks are always copied to the other node via the network.

 

[punkn123]

Hello Falk,

So, I don't have to break the cluster for this minor upgrade, right??

The upgrade from 8.1.3 to 8.2 includes Linux kernel updates?, which will require a system power cycle!.

Since this is a production cluster with client VMs, the only way to upgrade the nodes without affecting the VMs is to migrate them to another node right??

Also Is there any issue if some nodes remain on 8.1.3 while others are upgraded to 8.2? Will this cause any service disruptions, aside from missing out on the new features in 8.2??

Please let me know.

Thanks!

 

[Falk R.]

It is quite normal for clusters to be updated regularly.
You can use different versions during the upgrade, but it is recommended to update the version as soon as possible.

What kind of storage are you using?

 

[punkn123]

I have the following storages available in my VE:

Name                     Type     Status           Total            Used       Available        %
backup01                  pbs     active      1009829456         3893108       954566160    0.39%
BK-server                 pbs     active      4127328180        63254376      3854342276    1.53%
hdd                       rbd     active     44832884395        43890347     44788994048    0.10%
local                     dir     active     14884542464      1176144256     13708398208    7.90%
rmebackup01               pbs   inactive               0               0               0    0.00%
images                    nfs     active      2063105024         4741120      1953489920    0.23%
ssd                       rbd     active      7118127954        61032786      7057095168    0.86%
templa                    nfs     active       838110208       784617472        19850240   93.62%
zfs                   zfspool     active      4747952128            1536      4747950592    0.00%

Q) The upgrade from 8.1.3 to 8.2 includes Linux kernel updates?

 

[Falk R.]

Yes, the kernel will be updated from 6.5 to 6.8.

You have RBD pools, that is Ceph storage. You can migrate the VMs on the RBD pools between the nodes without downtime (if the CPU fits).
If you have activated HA for the VMs and set the shutdown policy in the cluster to migrate, then the node automatically migrates the VMs to other nodes on reboot and retrieves the VMs afterwards.
After the reboots, always wait until Ceph reports that everything is synchronized.

 

[punkn123]

Thanks for the assistance! Your response is greatly appreciated!

 

[punkn123]

Hello @Falk R.

We recently started upgrading our Proxmox servers, and after updating to the latest kernels (6.8.12.1 and 6.5.13.6), the system fails to boot and shows a kernel panic error. Attached a screenshot of the error. We reverted to the previous kernel version (6.5.11.7).

To utilize the new kernels, could this issue be related to disk configuration, or is there something specific we need to do before rebooting to ensure a successful boot? Any advice or guidance would be greatly appreciated.

Thanks!

 

[Falk R.]

Hello @Falk R.

We recently started upgrading our Proxmox servers, and after updating to the latest kernels (6.8.12.1 and 6.5.13.6), the system fails to boot and shows a kernel panic error. Attached a screenshot of the error. We reverted to the previous kernel version (6.5.11.7).

To utilize the new kernels, could this issue be related to disk configuration, or is there something specific we need to do before rebooting to ensure a successful boot? Any advice or guidance would be greatly appreciated.

Thanks!

The screenshot does not give me a direct indication. Is the firmware of all components on the old Dell R620 up to date? These are often problems with new drivers and old firmware.