Search results

Interesting. I'll consider it a false alarm then hahaha... I was kinda looking forward to tinkering.

I have a Ryzen 7 3700X - which seems to say it supports SEV: flags : ..... sev sev_es $ cat /run/qemu-server/host-hw-capabilities.json { "amd-sev": { "cbitpos": 47, "reduced-phys-bits": 5, "sev-support": true, "sev-support-es": true, "sev-support-snp": false } } When I try to...

If you're gunna be running Proxmox on 48 nodes, use your commercial support to lodge a ticket to get exact advice on your setup. You do have a support subscription given that Proxmox VE is core to your operation, right?

Thanks. And yes, for the sake of anyone coming across this later on and skim reading, DO NOT do this unless there is a functional access restriction that limits the world from using your system - or it'll make you an open relay to everyone and be a massive abusive target.

Had a chance to take a look at this today. I've done modifications to templates before, so not a huge learning curve here - but on the postfix config side, if I'm understanding it right, I just have to add `-o mynetworks=0.0.0.0/0` to the section for pmg.mail.int_port (which is port 26 in my...

Hi all, So I have a mate that I'm relaying email via my PMG instance. He's on a dynamic IP - which due to DNS updates works fine for sending mail TO his mail server, but having mail sent via PMG port 26 becomes difficult because I can't add a dynamic IP to the list of trusted networks. Given I...

Ahhh gotcha. Ok, i'll remove the fstab entry for /boot/efi and see how that goes.

Thanks for the tip - I did manage to find a missing security group on one of the VMs I imported from elsewhere. From looking at the firewall on each VM, the UI didn't indicate in anyway that the security group was missing. It was the only problem I managed to locate. I seem to remember that...

@shanreich - I have noticed an issue with the nftables firewall where this rule seems to be repeated thousands of times: meta protocol != arp ct state vmap { invalid : jump invalid-conntrack, established : accept, related : accept } Some logging shows that there might be a problem here with...

*nudge*

I've just been playing with my newly installed PVE 9.0.6 install - and when implementing BGP on the host using FRR, I noticed that if I shut down all VMs on the host, then the bridges will end up in the `state DOWN` state when viewing the output of `ip addr`. I managed to find a bit of a...

Since removing grub and using systemd-bootd via proxmox-boot-tool, now updates via apt show: root@mel-dc:~# apt dist-upgrade Summary: Upgrading: 0, Installing: 0, Removing: 0, Not Upgrading: 0 1 not fully installed or removed. Space needed: 0 B / 39.4 GB available Continue? [Y/n] y...

Interestingly, I left the firewall rules as they were and then enabled nftables - and all the created rules were there as expected. I added a macro rule, and saw its output just fine. It was almost instant. Interestingly, when I wrote this post, I had waited several minutes for the rule to...

Hi all, I've just done a fresh install of PVE 9.0.6, and noticed that when I add macro rules to a guests firewall - such as the HTTP / HTTPS macros, I can't seem to locate any rule that is created in the nftables firewall. I'm checking with `nft list ruleset` - and the guest chain doesn't seem...

https://bugzilla.proxmox.com/

Fixed in 4.0.11-4.

Hmmm - I still see this with 4.0.11.. The journal logs for the entire login to logged out are: Aug 07 06:56:55 pbs unix_chkpwd[829]: password check failed for user (root) Aug 07 06:56:55 pbs proxmox-backup-api[615]: pam_unix(proxmox-backup-auth:auth): authentication failure; logname= uid=0...

It occurs in 4.0.10 - even after reboots etc. I can't see a 4.0.11 anywhere at the moment. Also, I can login ok, but normally get logged out again quickly. The journal seems to show the auth failures, and the login session will be invalidated within a couple of page loads.

I'm using 1Password + the browser plugin to handle all passkeys.

I'm using Edge on linux (Fedora 42). No customisation of PAM, and only minimal changes to the underlying OS. I don't seem to see any errors for /api2/* in the journal when using a PBS realm user. EDIT: After adding a 2FA WebAuthn device to the PBS realm user, I logged out, then tried to log...