Search results

Hi guys. I have upgraded my cluster from PM6 to PM7.3-4. Versioning: pve-manager/7.3-4/d69b70d4 and Linux 5.15.83-1-pve. I observed that live migration does work only from older CPUs (like E5-2690 v2) to new ones (like Gold 6154) but not the other way around. When migration fails, the VM...

Sometimes we are just lucky. On another server it seems that when installing EFI partitions were also created: root@p38:~# proxmox-boot-tool status Re-executing '/usr/sbin/proxmox-boot-tool' in new private mount namespace.. System currently booted with legacy bios F32B-BD2F is configured with...

Thank you for your reply Fiona. We have 0.1 IO wait on PM hypervizor where VMs run. I did shutdown a VM, to destroy dirty map and ran backup job. It completed perfectly. We can not reproduce the issue anymore and all jobs are enabled yet again. Seems the issue was only on the first backup job...

Ah, tnx for the answer Stoikov. Are there any technical limitations to not be able to populate EFI partition when the system is booted via BIOS MBR? Maybe implement a command line switch for proxmox-boot-tool to call the EFI part of its code? I guess I could just take that EFI folder from one...

Thank you for you answer Fiona. High IO wait on backup server or on client (PM) server? On the PM / client / hypervisor there is zero IO wait. On storage there is. I will follow the link you gave me. I also found out that now that only dirty maps are used and data to transfer is below 30 - 60...

Hi, on a server where PBS 1 was installed first, then PM 6, i want to change from bios boot to EFI, so I can finally enable ZFS features. I have proper partitions and everything, but then, when I run proxmox-boot-tool, it just does not create EFI files on EFI partition. How to force...

Hi guys, I have had a cluster of ProxMox on version 6 as well as PBS installed quite some time ago. We use ZFS for everything. I recently added new nodes and upgraded everything,.. PM to 7 latest as well as PBS. When ProxMox nodes were upgraded, there were no issues. As soon as I upgraded...

Hi guys, I just got some Samsung PM943 1.92 TB NVMe drives. Will install latest PM on them with ZFS in RAID 10. Now I wonder, should I use ashift 13 when creating ZFS pool and then 8k block size when defining storage in PM or..? I guess people can answer faster than I can do benchmarks, ...

It crashed again after two months. So i decided and limited max_arc to 2 GB (from 4 GB as was default). Will see..

Hi @LnxBil , thank you for your reply and sorry for my late response (life happenz). To answer further, we already use KSM. I guess the solution here would be to restrict max ZFS ARC < ( SUM(VM RAM) + 2GB for PM host).

bump

Hi, after conversion of LXC to KVM which works, i have black screen / VNC console until I switch to tty1 with ctrl+alt+f1. Any ideas how to fix this, so that KVM 16.04 ubuntu based will start at tty1? I also just noticed, that when I use console for text input, it does not work. Picture shows...

Sorry for the late reply and thank you for the answer. Below linked is current arc_summary. Not from the time of the crash, however, hypervisor has not been rebooted since VM was killed. If you want some other data, let me know. Have had to link, post character limit was reached...

Hi, I just hit OOM killer today, due to ARC not retreating fast enough, so I guess it is still a problem nowdays. I have had SWAP on the same host and was almost empty when OOM killer hit. (I use MDADM SW RAID for it, cause ZFS has problems holding SWAP.) It;s sunday today, so I will continue...

I have had no OOM kills till now, even when I let ARC grow and it has to shrink due to memory pressure from VMs. Still cautious about doing it on all the servers.

Hi guys, I understand and always join an empty node to the cluster. But now I want to create cluster from one server which already contains VMs and is in production. Can the first node, the one you create cluster on, also contain VMs? If not, then I have to install new node, create a cluster...

Hi guys, does PM 6 work with PBS 2? I wonder if I should upgrade PBS from 1.1 to 2.*. It is used bt PM 6 cluster, fully updated. Have no plans for upgrade PM 6 to 7 a.t.m.

I guess I will have to fix the script so that it does not check all services. Till now if a service was listed in PVE, it was supposed to be running. But not anymore, so I guess this has to be changed: " def services return unless @options[:mode] == 'services' http_connect(path...

It did not help, it is still listed there. Maybe if remove systemd timed unit file as well?

So can i just remove sytemd-timesynced form the system and it will dissapear here as well?