Proxmox 8: New features and Upgrade process

[@annix493]

Every time I have one of your videos hit my feed, I consider it a sincere blessing. There is no one else out there who is so skilled at communicating complex ideas with such effortless clarity and detail. God tier tutorials.

[@mbunds]

Could not say it any better, thanks!

[@mandeepmails]

Ditto !

[@sprtwlf9314]

Dude your proxmox vids are the best around. And i was finally able to get pfsense working from your video on that. The physical diagrams you did at the beginning with the different hardware was super helpful. Youre a great teacher.

[@allanzfs]

This man is really smart and makes factual based videos.

[@mentalplayground]

I'm still on 7.x and will follow your advise. No rush with upgrade of 7 to 8.

[@joeyjojojr.shabadoo915]

You have earned my Sub. Your knowledge of Proxmox and the other topics that you make videos on are content that I can finally relate to, without breaking the bank on hardware that I can not afford or even have a use for. Keep up the good work.

[@gregorybehm8558]

Went smooth for me. Thanks for posting the video.

[@BillCClinton]

Tried an in place and it trashed my 7.4 server. Tried a clean install on the same computer that was running 7.4 just fine and am unable to map an SMB share to my unraid ISO folder. It maps just fine with 7.4.I get the following (create storage failed: storage 'unraid' is not online (500) even though it finds the smb shares just fine and works in 7.4 I'll keep tinkering with it. Thanks for the great videos! You rock my man!

[@BillCClinton]

A fix in case anyone else is having trouble mapping SMB share with proxmox 8 - Workaround: Input everything but deselect the Enable box. The storage than saves. Then edit it and enable it and it works...

[@AdrianuX1985]

They finally added a feature that has long been standard in pure debian, namely the "TUI" installer.

[@yasirrakhurrafat1142]

Not on all debians, but I get ya.

[@mpxz999]

Thank you for your help! You're the best!

[@CreamPolo]

Upgrade is super simple. My setup is very simple as well, just bunch of VMs. Upgrade went with no issues at all.

[@djmatropolis1]

Your tutorials are the best. Just finished with no issues and everything worked as you described. Appreciate the video.

[@jhon614]

I just finished upgrading using this guide and everything went perfectly. Thank you!!

[@VTOLfreak]

You might want to add a warning for people using ZFS on root, systemd-boot is not used by default anymore so you need to use apt to install it before you reboot. If you forget and reboot the system, the box won't come back up anymore. Note that this is mentioned on the upgrade page.

[@Jay2k1]

Are you sure? The upgrade page explicitly states: > Systems which have ZFS on root and boot in UEFI mode will need to manually install it if they need to initialize a new ESP (...). > > *Note that the system remains bootable even without the package installed.*

[@jhon614]

I just upgraded from proxmox 7 to 8 using this video, and I didn't do anything special to reinstall systemd-boot, and I have the boot drive configured as a zfs mirror. My system came up just fine. Do I need to do anything else? Did I just get lucky because it's a zfs mirror and not raidz?

[@ElectronicsWizardry]

@jhon614 if your using uefi boot and zfs for the boot drive you should install systemd-boot package. It looks like the package is needed to setup the efi partition so its likely the system will boot without by may have issues with future upgrades or changing a boot drive in the mirror. I’d install the package in your case. It looks like this is included in the pve7to8 script so you may have gotten a warning.

[@jhon614]

@@ElectronicsWizardry done. Thank you so much!

[@rklrkl64]

One thing that was missed here is how well upgrading a Proxmox cluster from 7 to 8 goes. In particular, live migration between consecutive major versions has been broken in some previous Proxmox releases, which I find infuriating because it means you have to offline migrate when bringing VMs back onto an upgraded node, causing VM downtime. Proxmox devs just shrug their shoulders when tnat happens, but really they should ensure that live migration from the last version of X.* (7.4 in this case) to X+1.0 (e.g. 8.0) works OK. Another thing that's quite tricky in a Proxmox cluster is if you mix Intel and AMD x86-64 servers - this isn't far-fetched because there may be a great offer on, say, AMD servers wihile you're about to decommission part of your existing Intel-only Proxmox cluster, but you can't afford to replace all the Intel servers. I'd like to see some effort made by Proxmox to document/code for the best way to handle mixed Intel/AMD clusters (I know they don't recommend such a mix, but reality is that it can happen, especially as AMD are seeing increased server market share ever since Epyc came out). I'd like there to be a set of CPU flag aliases that are for every Intel and AMD server generation for the past, say, 6 years and then a crosstable on a Wiki page showing the combinations that can safely live migrate in both directions without virtual CPU failures. You then look at the crosstable to find the oldest generation of Intel and AMD you have in your cluster and set the CPU flags for all your VMs to that alias in the crosstable, ensuring you can live migrate from any node to any other mode without issue.

[@jttech44]

I mean, Debian 11 isn't EOL, so, unless you *need* Debian 12/PVE8 right this moment, you've got plenty of time to wait and plan for that downtime. And I'd suggest giving it a few months minimum before even testing this seriously. It's likely to go through some fairly major patching as Debian 12 gets rolled out.

[@kirksteinklauber260]

Thanks for the tutorial and guidance!!! It will help me in my journey to test and update my servers!

[@5speedfatty]

thanks for the feature review, I didnt even know 8 had been released.

[@saidneder41]

The best videos ever brother god tier tutorials!!

[@allanzfs]

I wish that the developers would spend more time creating a graphical interface that would provide a clean method of working through the top tasks without resorting to a series of shell commands to make them work. For example, creating a cluster is straightforward though deleting the cluster is far more complicated and involves shell commands. Perhaps your viewers could create a top-20 wish list so the usage is of the same level as Esxi?

[@jttech44]

It's linux, some things are just easier in the shell. I'd much rather the UI stay big picture and leave the uncommon configs out.

[@mtruo001]

Hey just want to say im g,ad you are posting more videos. Your content is great and more importantly there is nothing else like it out there. Speaking about the knowledge shared in an effective way.

[@TonyGonzales]

Love your videos man, the explanations are on point.

[@ierosgr]

At 2.24 you mention about the default cpu type being the QEMU64. That one isn t the default. KVM64 is instead .

[@JonatanCastro-secondary]

I did an in-place upgrade from 7.4 to 8. Everything went okay! :phew:

[@darthkielbasa]

EW has to be one of the most undersubscribed gems in existence.

[@alexthebear]

Thanks, man! I just updated mine. Works nice.

[@bdhaliwal24]

Thank you! Very useful guidance

[@Rood67]

I figured out how this guy is so smart with ProxMox; he’s Riff Raff from Rocky Horror and has enhanced, extraterrestrial knowledge we mere mortals would never understand.

[@tjoleary8738]

Prepping for upgrade.

[@danielhansson9810]

Proxmox should hire you.

[@HarryComa]

I see you still have the SUN 2u machine around. I bought one after your first video on it a few years back, now i have two. Love them, soo much better than DL380 G8 and similar.

[@greenftechn]

Nicely done, thank you!

[@StevenLynn]

First physical host of mine I tried upgrading would boot. It would get to the mount drive section on screen, then "hang". Give the host another minute and everything would function as normal (website, vms, networking). Just the login prompt never appeared nor the other TTY's. Ended up just fresh installing all four of my hosts as a clean slate. Other than the boot issue 8.0 has been rock solid and all my VM's have been functioning as they did for months on 7.x. Keep in mind some of my VMs have gone from VMware 6.7 to XCP-ng 8.x to Proxmox 7.3->7.4 now 8.0 in just this year. :)

[@danieldewindt3919]

Excellent

[@foureight84]

I was really hoping they would add support for serial console for the installer by now. Its quite annoying having to install debian first and then the proxmox ve packages on serial console-only servers.

[@paullee107]

I hate how youtube offers no way to message a content creator. I have a question about proxmox (and PBS) and backing up non-PVE VMs on the PBS... can I contact you?

[@Zarathustra-H-]

Lol. I have upgraded in place since Proxmox VE 4.1, so I've never seen a GUI installer. As I recall the installer for 4.1 was text as well.

[@rowenarrow]

Great video!

[@RyanKnowsTechStuff]

is there a discord for the community so we can chat with others about our virtualization tech issues?

[@edgecrush3r]

Moving to this new kernel, i hope vGPU is still possible.

[@RemyDMarquis]

Thank you

[@keivankarimi]

Nice video. I also followed some other tutorials from you and it was pretty helpful. I noticed that Proxmox 8 is handling PCI pass through for IGPU in a different way than Proxmox7. Could you please make a video on that?

[@keivankarimi]

I forgot to say tnx :)

[@MauroSedrani]

Great video. Thanks!

[@jjones503]

Helpful video. Thanks

[@viruslab1]

I have updated it in place yesterday. works like a clock.

[@i420bdawg]

one thing to mention that while it is in the docs is that if you have an Nvidia GPU working on Proxmox 7.4 it is not yet supported on Debian 12 / Proxmox 8. I am having to go back and wipe my proxmox install then reinstall because of this. I wish they had a way to revert back (Downgrade) to Proxmox 7.4

[@ElectronicsWizardry]

How are you using the gpu? Is this just for passthrough or are you using it for passthrough, or is proxmox managing the vm.

[@thenanook]

Smart dude, subscribed

[@miles267]

Hi @electronicswizardy can you please confirm how to test whether AES-NI is enabled in a Proxmox VM? For example Untangle or pfSense? I'm trying to run grep commands but they're not returning anything. When I run these 2 commands, the results are almost identical. Also wasn't sure whether some sort of flag had to be set in the guest VM (in my case Untangle) to enable. In my case, I don't want to have to reinstall the untangle VM if I can avoid it. openssl speed -elapsed aes-128-cbc openssl speed -elapsed -evp aes-128-cbc Thank you very much for your insight.

[@cbaservs]

my ceph dashboard page is not working properly anymore after the update...cannot reach the 8443 page anymore

[@doutu4624]

Hey, dude! Seeing your hairstyle, I can tell you're really smart!

[@fredwupkensoppel8949]

Well, mine has Dropbear in the initrd for unlocking the FDE and GPU pass through configured - something's telling me it's gonna break. lol

[@subrezon]

Now just waiting for OpenWRT 23.05 to release, and then upgrade everything in my server rack at once.

[@TheMongolPrime]

I tried doing a fresh install of 8.0.2, and ran into installer issues with nvidia drivers. There are some forums posts with more details, but basically you have to remove your nvidia cards before installing fresh, then reinstall them. I can't do that right now, so I'm waiting for a patch or update to fix that driver issue.

[@jttech44]

Just disable them in the bios, then re-enable them once you're installed.

[@wierdo1111x]

Blacklisting the Nvidia drivers worked for me.

[@andresrocha9579]

How does affect an upgrade in one of the server to the proxmox cluster?? (i have 3 servers with 7.2-3 version) Thanks a lot!

[@ElectronicsWizardry]

You should be able to just upgrade each node at a time. I’d move all the VMs and containers off a node first. The. Update. Then repeat for all nodes in the cluster.

[@andresrocha9579]

@@ElectronicsWizardry I will do that, thanks you very much! You are awesom

[@jack44829]

Hi , I really need help how to migrate my old physical server which is in centos Linux on rconfig to the new Proxmox server..can you assist on this and I will appreciate on it

[@ElectronicsWizardry]

I go over how to migrate from a physical server to Proxmox in this video I made earlier. kzbin.info/www/bejne/apezXpyidtSSgdU. Let me know if you have other questions.

[@morto360]

Would the backup//restore option also work on older systems like Proxmox 5.x to 8?

[@ElectronicsWizardry]

I believe the backup format form v5 till now isn’t changed so it should work fine. But it might be best to install proxmox 8 in a vm or on a spare system to make sure it can read the back up file.

[@phsouzabr]

Did an in-place upgrade and it borked the server. Gonna have to do a fresh install and restore now =D

[@ElectronicsWizardry]

I'm curious, do you know what broke or what the issue was?

[@ezteok1]

Nice tutorial! Does this upgrade process goes smooth for a cluster setup as well?

[@ElectronicsWizardry]

This process should work with clusters with a few changes. I’d upgrade the cluster one node at a time. First move all the VMs off a node or shut them down. Then upgrade and reboot the node and repeat for all the nodes in the cluster.

[@ezteok1]

@@ElectronicsWizardry Thanks, I will give it a try. I have good backup's so if things really go wrong I can always start from scratch and restore my backups after the upgrade. Thanks for your advice!

[@jttech44]

@@ezteok1 Do yourself a big favor and test your restore process before you need it.

[@ezteok1]

@@jttech44 thanks! But i test it many times. I realy like proxmox backup (and restore), very good and easy to use. I was more concernd about my shared nfs storage (no vms whatsoever without it), , raspberry pi qdevice, etc.)

[@ezteok1]

@@ElectronicsWizardry did it, upgrade from 7.3-4 to 7.4-15 to 8.0.3 on two nodes (the third node in my cluster is a raspberry Qdevice). One node is older than the other (different changes) and so I received different questions and notifications per node during the upgrade. Confusing but doable :) Thanks!

[@Stampe-OZ0AS]

In my stupidity, in an attempt to free up disk space, I have deleted all my "ISO Images" under "Local (pve) -> ISO Images. Is it possible to recreate these images? I have also not made a backup of the VM' s

[@ElectronicsWizardry]

Are these iso images things like windows and Linux installers? Generally I find those easy to get on public servers. I don’t think iso images are backed up with the vm so unless you have snapshots or anouther backup those are likely gone.

[@Stampe-OZ0AS]

@@ElectronicsWizardry Yes the was installers

[@Stampe-OZ0AS]

@@ElectronicsWizardry And thanks for your reply, I will try to find the files. The error it shows is: " TASK ERROR: volume 'local:iso/ubuntu-22.04.2-desktop-amd64.iso' does not exist"

[@shephusted2714]

aaaawwww yeah - please explore doing a cluster without proxmox too - you sort of don't need it, it does make it easier but it is not the only way, my main thought from this is - don't break debian; my only request to you is to make an ultimate cluster sort of along lines of old linux journal ultimate linux box articles - you can do it annually to gauge pace of change with hw/sw - right now it looks like ultimate proxmox node is probably 16 core/128gb ram/1tb nvme for os and 4tb for netfs and then 100g plus an optional gpu for compute and ai modeling and training - you could literally do this with a hp z440 - prices for very capable nodes will keep dropping... that is one thing we can count on going forward

[@jttech44]

More nodes = More better. I'm running 12 in prod currently, with another 3 on order. It really gives you a ton of flexibility to have extra compute hanging around.

[@wierdo1111x]

128GB of RAM is a ton for 16 cores...

[@jttech44]

@@wierdo1111x Depends on the application. If it's hosting simple databases, potentially not enough, if it's hosting file servers, likely overkill.

[@pepeshopping]

This is why Proxmox is NOT yet a Production ready solution! The day they get a ONE click. config backup/restore AND a ONE click version upgrade. Just like ESXi, TrueNAS, etc.

[@jttech44]

Currently running a 12 node 200TB cluster in prod, works great, very minimal headaches compared to ESXi (which is very picky about hardware choices and support generally sucks at solving anything remotely exotic). Having everything run on debian is really a huge boon when it comes to administration overhead.

[@TaT-YoK]

What is the 5U server in the back?

[@ElectronicsWizardry]

The one with the 3 drives in the middle is a 4u case. I currently have it filled with a 5950x as my main server.

[@TaT-YoK]

@@ElectronicsWizardry I'm really looking forward to your videos.

[@officiel216]

Give me the installation commands on the Debian server

[@Aplifiedproductions]

Are you a proxmox employee causse it alsmost as if you are with your knowlage base lol. Love the content

[@ElectronicsWizardry]

Not an employee, just like their products and making videos about them.

[@jeromealtariba7339]

main issue is that when running the first command to upgrade to 7.4, I have an error saying I'm running out of space to update the pat database. I don't know why suddendly I have not enough space in my root directory. I tried apt clean, and other commands to release space, but this failed. What are the other options ?

[@ElectronicsWizardry]

I'd try running commands like du -xa / | sort -n -r | head -n 20. This will search your boot drive for the 20 largest folders. You may have some some cached cache packages you can delete or tempt files. Its hard to know what the files are with out seeing the system myself. I also like tools like ncdu and gdu for this, but if your drive if full then can't be installed and used.

[@jeromealtariba7339]

@@ElectronicsWizardry Hi, thks for your reply. In fact, it was my backups that were saved in a root subfolder (I guess default one). Need to change this

[@pyrhockz]

❤

[@ManfredBartz]

Good tutorial - thanks. 👍 At kzbin.info/www/bejne/oqCxgZyjqs2ip8k a mention of "apt dist-upgrade" would have been useful, but the proxmox wiki covers that.

[@notoriouslycuriouswombat]

jonny lee miller if he was a real hacker after 25 years instead of a actor

[@allanzfs]

Made the upgrade from 7 to 8 and am really disapointed! Why? The upgrade was smooth, without any problems, but a Windows 10 would not start and an old Windows Server had numerous failure to run services errors. All worked well in Proxmox 7, but 8 is a fail. And a new instance of Windows 10 using a new installation ISO failed to even start the installation! All Linux VMs worked properly. So be careful.

[@ElectronicsWizardry]

I'm curious what the issue was that caused the VM not to start. Did the vm start in Proxmox? Did windows in the VM report errors?

[@talon32779]

Realtek R8169 driver in Proxmox 8 is buggy

[@randleqgod]

Stay away from Realtek. Too many driver issues.

[@Darkk6969]

I'd stick with Intel and Mellanox. Realtek just asking for headaches.

[@mathieuleclerc4136]

please send the url

[@cyklondx]

guy looks like golum

[@VusalMDadashzadeh]

find some barber homie....

[@ferencszabo3504]

Man, your example is f.king confusing, next time don't dance around theme

[@blakenorthrup]

Hey, getting the errors: E: Sub-process /usr/share/proxmox-ve/pve-apt-hook returned an error code (1) E: Failure running script /usr/share/proxmox-ve/pve-apt-hook Won't upgrade

[@MrEe1010]

When running pve7to8, I get "command not found". I also get errors: "E: Sub-process /usr/share/proxmox-ve/pve-apt-hook returned an error code (1)" and "E: Failure running script /usr/share/proxmox-ve/pve-apt-hook"

[@ElectronicsWizardry]

Check that you have updated your system recently as it has been added to the package repository recently. Also check that your proxmox repo is working correctly. By default you need to enter a license for the enterprise repository or add the no subscription repo.

[@MrEe1010]

@@ElectronicsWizardry Thank you. You were correct, I didn't have the repository working correctly. Now that it is fixed, I was able to upgrade to PVE8.

[@RomanShein1978]

Great video!