udo's latest activity
-
Uudo replied to the thread Ceph security issue CVE-2025-30156.of course it's not tivial - especialy if you have cephfs and radosgw running too (like me).
-
UI am on the Ceph users email list, and it seems the fix is not a simple update, and will not be trivial. You have to use new authentication keys, aes256k, and rotate the keys used, and for that (for rbd storage at least), shutdown all VMs. And...
-
Uudo posted the thread Ceph security issue CVE-2025-30156 in Proxmox VE: Installation and configuration.Hi, there is an critical security issue on ceph https://notcve.org/cve/CVE-2025-30156 I assume you are working on new packages. Can you estimate when they'll be ready? Best regards Udo
-
Uudo replied to the thread [SOLVED] Question about cpu-models and cpu flags.Hi, yes you are right! I took an older grml to test the cpu-flags. With the actual one (kernel 6.19.13) the lahf_lm flag was visible inside the VM on amd too. Best regards Udo
-
Uudo reacted to fiona's post in the thread [SOLVED] Question about cpu-models and cpu flags with
Like.
Hi, which kernel is running inside the guest? Maybe it's the issue fixed by this commit (contained in kernels from v6.15 onwards). -
Uudo replied to the thread [SOLVED] Question about cpu-models and cpu flags.Hello Arthur, of course… lahf-am is there: qemu-system-x86_64 -cpu help Available CPUs: 486 (alias configured by machine type) 486-v1 Broadwell (alias configured by machine type)...
-
Uudo replied to the thread [SOLVED] Question about cpu-models and cpu flags.Hi Arthur, thanks for the info. The lahf-lm flag is strange. Have testet a limitied cpu with "flags +aes;+lahf-lm" on an Intel-Node (E5-2640 v4) and the flag is visible inside the VM. Then I migrate the VM back to an AMD-Node (EPYC 9454P) and...
-
Uudo posted the thread [SOLVED] Question about cpu-models and cpu flags in Proxmox VE: Installation and configuration.Hi, we have cluster with mixed CPUs and we use as cpu-settings something like this (amd) cpu: x86-64-v3,flags=+ibpb;+virt-ssbd;+amd-ssbd to avoid to select the additional flags I tried to define an custom cpu type. Two things are not working...
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Found the issue - was my fault - from an prior installation on other hardware (without boss card), there was still "GRUB_CMDLINE_LINUX="root=ZFS=rpool/ROOT/pve-1 boot=zfs" in the grub-default-file… Thanks for comments!!
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Hi, perhaps that has something to do with the issue. On another similiar R6615 which I upgrade this morning, the packages are not installed. On the node with the issue, was systemd-boot removed, but systemd-boot-efi installed… dpkg -l | grep...
-
Uudo reacted to miguelsousa's post in the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk with Like.
Did you run apt remove systemd-boot on the upgrade? I'm planning a similar upgrade and have on my notes: apt install --reinstall grub-efi-amd64
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Hi, secure boot is disabled in the bios. Strange, then efibootmgr shows \EFI\Proxmox\shimx64.efi I updated another Node this morning, also on a R6615 with Boss-Raid-1-Card, which runs well. On this Node Secure Boot is also disabled and...
-
Uudo reacted to longer's post in the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk with Like.
If not already done, test to disable secure boot in BIOS and try if it can boot
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Hi, of course efibootmgr -v BootCurrent: 0001 Timeout: 0 seconds BootOrder: 0004,0003,0000,0001 Boot0000 Virtual Floppy PciRoot(0x3)/Pci(0x7,0x1)/Pci(0x0,0x4)/USB(0,0)/USB(3,0)/USB(0,0)/Unit(1) dp: 02 01 0c 00 d0 41 03 0a 03 00 00...
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Yes, I mean the same. The server aren't bootet and was not available on the network. Not with the interface for clustercommunication, nor with the openvswitch-bonds…
-
Uudo replied to the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk.Yes, the console message from iDRac is correct - the node aren't booted without display output.
-
Uudo posted the thread [SOLVED] after upgrade to PVE9 one Node stuck on Loading initial ramdisk in Proxmox VE: Installation and configuration.My fault! There was still an "root=ZFS=rpool/ROOT/pve-1 boot=zfs" in the default-grub file due an prior installation on other Hardware without boss-card. Hi, this mornig I've upgrade 3 Nodes of an Cluster to proxmox-ve: 9.2.0 (running kernel...