Proxmox VE 7.4 released!

[fiona]

Hi,

noVNC console doesn't work. VM is debian 11, VM is up and pingable.

Mar 31 10:40:04 proxmox-mon-01 pvedaemon[158819]: starting vnc proxy UPID:proxmox-mon-01:00026C63:0070E3BD:64269C64:vncproxy:101:root@pam:
Mar 31 10:40:04 proxmox-mon-01 pvedaemon[1304]: <root@pam> starting task UPID:proxmox-mon-01:00026C63:0070E3BD:64269C64:vncproxy:101:root@pam:
Mar 31 10:40:13 proxmox-mon-01 pveproxy[1310]: worker 154041 finished
Mar 31 10:40:13 proxmox-mon-01 pveproxy[1310]: starting 1 worker(s)
Mar 31 10:40:13 proxmox-mon-01 pveproxy[1310]: worker 158850 started
Mar 31 10:40:14 proxmox-mon-01 pvedaemon[158819]: connection timed out
Mar 31 10:40:14 proxmox-mon-01 pvedaemon[1304]: <root@pam> end task UPID:proxmox-mon-01:00026C63:0070E3BD:64269C64:vncproxy:101:root@pam: connection timed out
Mar 31 10:40:14 proxmox-mon-01 pveproxy[158849]: worker exit

the error message indicates that the vncproxy didn't get a connection for 10 seconds and thus timed out. Are you using the noVNC console in the web UI or are you using an external client? Please open the developer tools in your browser (usually Ctrl+Shift+C) and go to the Network tab there. Can you see any failing requests there when you try to open the noVNC console? Did you flush the browser cache after upgrading?

 

[czechsys]

Hi,

the error message indicates that the vncproxy didn't get a connection for 10 seconds and thus timed out. Are you using the noVNC console in the web UI or are you using an external client? Please open the developer tools in your browser (usually Ctrl+Shift+C) and go to the Network tab there. Can you see any failing requests there when you try to open the noVNC console? Did you flush the browser cache after upgrading?

Yep, the pve classic and special "flush browser cache after pve upgrade" problem as usually... Console works.

 

[ozdjh]

Hi all,

The release notes say that the option to disable the guest agent fsfreeze during a backup was included in 7.4 but I can't find it. Is that exposed in the UI? It doesn't show up on the VM Options page.

 

[fiona]

Hi,

Hi all,

The release notes say that the option to disable the guest agent fsfreeze during a backup was included in 7.4 but I can't find it. Is that exposed in the UI? It doesn't show up on the VM Options page.

AFAIK, the GUI patches were not yet applied. But you can set the option via CLI/API, for example qm set 149 --agent enabled=1,freeze-fs-on-backup=0,fstrim_cloned_disks=1. Note that you have to pass the full settings, it's not possible to touch the freeze-fs-on-backup part only (i.e. freeze-fs-on-backup would be reset to its default 0 if it is not specified). The documentation is here: https://pve.proxmox.com/pve-docs/chapter-qm.html#qm_qemu_agent

 

[ozdjh]

AFAIK, the GUI patches were not yet applied.

Oh, that's a shame. We've been waiting on that feature so we can look at moving to PBS from our own backup solution. We'll have a look at this on our lab cluster and wait for the UI to catch up with the vzdump change.

 

[uebmaster]

Try running pveam update. Any error messages?

You are great!!!

Thanks, it works.

 

[susie]

hey very excited on new release especially Dark Mode looks fantastic. Hope my 7950x cpu issue will get resolved

 

[mr44er]

Two nodes crashed during the update (from Linux 6.1.0-1-pve #1 SMP PREEMPT_DYNAMIC PVE 6.1.0-1 pve-manager/7.3-3/c3928077 ) with hard reset. They had 80+ days uptime, maybe next time I'll give them a reboot before updating.
Anyway, after a dpkg --configure -a && apt update && apt dist-upgrade everything smooth! Nice work and thanks!

 

[michaelj]

Hi,

What is the EOL (security) for versions 7.X please ?

 

[shanreich]

Hi,

What is the EOL (security) for versions 7.X please ?

Currently tba, it depends on how long bullseye (Debian 11) stays oldstable after the release of bookworm (Debian 12). This usually lasts around one year after the initial release of the new version. See our FAQ [1] and the respective Debian Wiki page [2] for more information.

Please note that using the latest version is always recommended and preferable over running older major versions.

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_frequently_asked_questions_2
[2] https://wiki.debian.org/DebianOldStable

 

[gsmitt]

Just start using this software to replace Hyper-v. So far it really nice. Easy to install. Been testing for a couple days. Going to put heavy load next week on a cluster. We actually installed Prox on Beelink ( mini-server 32Gb ram, 8 cores & 500GB drive). I was able to run 11 VM's with 4*4 100Gb drive no issues. You got to like thin provisioning

 

[pumo]

is upgrading 7.3-6 to 7.4 allready safe ? I havent upgrading since 7.4 was released. I know if it works dont fixit

 

[Dunuin]

is upgrading 7.3-6 to 7.4 allready safe ? I havent upgrading since 7.4 was released. I know if it works dont fixit

Depends on what you are doing with it. Would be bad running some virtualized Webservers without installing any security patches.
And yes, from what I've read in these forums 7.4 isn't more problematic than 7.3.
If you care that much about stability you might want to spend some money on any of the subscriptions so you can make use of the better tested enterprise repo.

 

[chrcoluk]

Should i remove 6.1 and 5.19 kernels before installing this update then, as you state its 5.15 or 6.2 only?

 

[t.lamprecht]

Should i remove 6.1 and 5.19 kernels before installing this update then, as you state its 5.15 or 6.2 only?

Could be also done afterwards, but it may save some space and depending on how you boot that could be important.

For example, if you boot with ZFS as root and have used a recent ISO to install that (or boot EFI) then your boot partition is most likely 512 MB big, and thus you can only have _that_ many kernels on there.

So, removing the meta packages, e.g. via: apt remove pve-kernel-5.19 pve-kernel-6.1
And then auto-removing the respective kernels, e.g. via: apt autoremove
does ensure that you have as much space as possible without removing any kernel that is currently in use.
You can always re-install those kernels later, whyever that'd be needed.

 

[dirk.h]

• A fully-integrated "Proxmox Dark" theme variant
• Sorting of resource tree by name or VMID
• Q: Can I upgrade Proxmox VE 6.4 to 7.4 with apt? A: Yes

Those "small" things really means a lot to me. Thank you!

As a tiny feature Request: Sorting by reversed FQDN (.com.proxmox.forum), or regular written (forum.proxmox.com), but sorted by the domain level (tld.sld....)

 

[Winet.maier]

We just did an UpDate from 7.3 to 7.4 and ran into the following problem. VM with CPU devinition AMD EPYC Rome could not be migrated to the just updated hypervisor. After a migration to AMD EPYC (without Rome) the whole thing worked. It's no big deal, but all hypervisors have the same CPU's, AMD EPYC Rome. When all VM's were migrated, we could migrate again with AMD EPYC Rome normally. This is a very strange error that should not occur. I hope we can find an explanation.

 

[davemcl]

We just did an UpDate from 7.3 to 7.4 and ran into the following problem. VM with CPU devinition AMD EPYC Rome could not be migrated to the just updated hypervisor. After a migration to AMD EPYC (without Rome) the whole thing worked. It's no big deal, but all hypervisors have the same CPU's, AMD EPYC Rome. When all VM's were migrated, we could migrate again with AMD EPYC Rome normally. This is a very strange error that should not occur. I hope we can find an explanation.

This might be your issue.
https://forum.proxmox.com/threads/cant-use-epyc-rome-cpu-after-update.125336/#post-547510

 

[spirit]

We just did an UpDate from 7.3 to 7.4 and ran into the following problem. VM with CPU devinition AMD EPYC Rome could not be migrated to the just updated hypervisor. After a migration to AMD EPYC (without Rome) the whole thing worked. It's no big deal, but all hypervisors have the same CPU's, AMD EPYC Rome. When all VM's were migrated, we could migrate again with AMD EPYC Rome normally. This is a very strange error that should not occur. I hope we can find an explanation.

same rome model/frequency ? same microcode version ?

did you have any error on migration task ?

I known that amd have remove xsaves from last microcode because it was buggy, but I think it's still in the qemu rome model definition. (a new model -vX is avaiable in new qemu)

 

[Winet.maier]

This was during migration and I dont remember the error. But the TEmplate was still AMD EPIC Rome and if we make a new VM from it it says:
kvm: warning: host doesn't support requested feature: CPUID.0DH:EAX.xsaves [bit 3]
kvm: Host doesn't support requested features
TASK ERROR: start failed: QEMU exited with code 1