Search results

  1. K

    LXC load averages are wrong in Zabbix on PVE 8.x in LXC containers

    After upgrading to PVE 8 and several CTs to Debian 12, I experience a loadavg discrepancy. The load averages shown in the CTs are that of the host, divided by the number of CPU cores of the LXC container (per-core loadavg in Zabbix), instead of the load of the LXC guest itself. The commands...
  2. K

    [SOLVED] Moving cluster operation traffic to another network

    Well, for whomever finds this thread among the other similar ones, the issue was solved by fixing /etc/hosts to align the host names with the desired subnet, and then restarting pve-cluster.service and then corosync.service. Replication and other heavy traffic is flowing on the network it was...
  3. K

    [SOLVED] Moving cluster operation traffic to another network

    I made a mistake when I created a 5-node cluster. I added all nodes by IP to the cluster and `pvecm status` now shows: Cluster information ------------------- Name: cluster2 Config Version: 5 Transport: knet Secure auth: on Quorum information ------------------ Date...
  4. K

    [SOLVED] Azure to PVE migration problems

    OK, it was actually a pretty lame issue. I had to clean up the garbage in GRUB_CMDLINE_LINUX in /etc/default/grub and presto, booting like it should. I leave this "solution" here for search engines.
  5. K

    [SOLVED] Azure to PVE migration problems

    I'm struggling with a seemingly simple task. I've copied the contents of an old Debian 8/Jessie VM from Azure to a PVE 6.1-7 host using rsync. I've set up boot, etc., and the VM does start booting, but I see the messages per attached image and it's stuck there forever. The root partition is a...
  6. K

    [SOLVED] Weird issue with iptables on PVE 6.3 host - rules are in the kernel but ineffective

    I've upgraded a 5.2 system to latest 6.3 and experiencing something I've never before. It might or might not be PVE related, I have no idea. The gist of it, is I have a custom iptables firewall that loads without problems and does work. I had to allow a new IP for admin access, something like...
  7. K

    LXC swap size change/regression between PVE 6.3 minor versions

    Apparently, this is the way it should be with cgroup v1. Cgroup v2 is supposed to show swap separately again but it's not active in PVE right now. It's unexplained why it was shown correctly (separately, as it should be in v2) in earlier 6.x versions. It's a mess, anyway. See also here.
  8. K

    LXC update breaking up write permission

    It seems lxc-pve 4.0.6-2 is in the no-subs repo, so instead of downgrading it should be enough to upgrade to latest... Though I couldn't try it, yet.
  9. K

    LXC update breaking up write permission

    Not now, as the system is live and people started working on it, but as an idea, I might try later. But i was under the impression that nesting is only necessary when processes try to create security namespaces and other security measures using cgroup inside CTs so my gut feeling says it won't help.
  10. K

    LXC update breaking up write permission

    It has nothing to do with the problem at hand. It's just the tun device config, which is working fine. There's a small, but very definite difference in the CT vs. the older version of pve-lxc. The following mount entry is not present in the newer version: proc on /proc/sys/net type proc...
  11. K

    LXC update breaking up write permission

    Sure, here it is: # pct config 101 arch: amd64 cores: 1 hostname: vpn.xxx.private memory: 512 net0: name=eth0,bridge=vmbr0,hwaddr=8A:D3:78:C4:11:22,type=veth onboot: 1 ostype: debian rootfs: thinhdd:vm-101-disk-1,size=8G swap: 128 lxc.cgroup.devices.allow: c 10:200 rwm lxc.mount.entry...
  12. K

    LXC update breaking up write permission

    Same issue here with a CT running OpenVPN. It needs to enable IP forwarding in the kernel (e.g. /proc/sys/net/ipv4/ip_forward). Fails with write permission error. Quite the issue, it kills most network/routing oriented containers. Will try downgrading in the meantime. Downgrading...
  13. K

    LXC swap size change/regression between PVE 6.3 minor versions

    I've just upgraded a server to PVE 6.3 (pve-manager/6.3-3/eee5f901 (running kernel: 5.4.78-2-pve)) and see a sudden change in how LXC swap amount is presented. For a long time in older versions in LXC (maybe starting with 4.x) the swap size was shown and calculated as RAM+swap, but thankfully...
  14. K

    SSD Wear Out Calc?

    @ThinkAgain There are 2 more factors to take into account. First, you seem to use ZFS where commits from the ZIL cause double writes on the pool (not always; it's complicated). The Total_LBAs_Written attribute is total host writes, not taking write amplification into account (double writes are...
  15. K

    ZFS: raid disks

    It's normal, because part1-2 is used for EFI systeam partition and boot, respectively (if I remember correctly). And it's better to have a partition table anyway to avoid "accidents" with bare disks.
  16. K

    pveproxy not listening on IPv6, only IPv4

    SLAAC provides a dynamic address. At least it used to, in that environment. Now they seem to provide a dynamically assigned, but static EUI-64 address for the servers (makes a lot more sense, honestly). I can add that to the hosts file, but see my other concerns. Plus I see the whole fiddling...
  17. K

    pveproxy not listening on IPv6, only IPv4

    The systems are dual stack, not IPv6-only. If I add the v6 address as well, won't it create confusion in the cluster communication and other things? On the other hand, the HNs use dynamic v6 addresses via SLAAC. This is a technical limitation in that system. So I can't just add an address...
  18. K

    pveproxy not listening on IPv6, only IPv4

    I'd like to access my PVE servers on their IPv6 addresses in a DC with native IPv6 available. Currently it's impossible as the pveproxy service is not listening on IPv6. The systems have proper dual stack setup, default (high) IPv6 preferences. Could it be fixed so the pveproxy service listens...
  19. K

    Separate loadavg for individual containers

    Yeah, my bad, it's a separate package. And yes, my concern is also is that people seem to be infatuated with load values and make (mostly false) assumptions based on them.

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!