UdoB's latest activity

  • UdoB
    UdoB replied to the thread PBS with NFS.
    The physical storage in the NAS is solid state or rotating rust? Topology and filesystem? Keep in mind that PBS needs "some" IOPS. My very first PBS was classic HDD only. It worked. But after some time, after storing some amount of data, it got...
  • UdoB
    UdoB reacted to Johannes S's post in the thread Proxmox CPU temp and Using in VMs with Like Like.
    Nope it's about priorities and return of investment (that's my guess at least). I'm willing to bet my homelab hardware that Proxmox Server Solutions GmbH (the company behind PVE, PBS and PMG) generates most of it's income not from us homelabbers...
  • UdoB
    Was wieder mal ein tolles Beispiel dafür ist, warum man von diesen Skripten die Finger lassen sollte, wenn man nicht weiß, wie die funktionieren. Darum: Danke, in Zukunft werde ich hierdrauf verweisen, wenn ich ein Beispiel dafür brauche ;)
  • UdoB
    UdoB reacted to news's post in the thread Lahme Kiste with Like Like.
    Nun ja HDDs und ZFS funktioniert, aber nur, wenn man noch einen ZFS special device über 2x SSDs im zfs mirror mit einbindet. Bei ZFS geht es immer um die IOPs für random 4k R/W, die kann man massiv aufblähen, wenn man die ZFS Metadaten dann in...
  • UdoB
    UdoB reacted to SteveITS's post in the thread CEPH LICENCE with Like Like.
    @zoher chelghoum You’ll need to change repositories to no-subscription to get updates.
  • UdoB
    UdoB reacted to Johannes S's post in the thread How to backup a data storage (paranoid) with Like Like.
    Why don't you use a PBS sync job? Shouldn't this be more robust than than rsync?
  • UdoB
    UdoB reacted to spirit's post in the thread Cluster Fuckery with Like Like.
    when you edit /etc/pve/corosync.conf (in normal conditition, when you have quorum and write access to /etc/pve), the file is copied locally to each node in /etc/corosync/corosync.conf , then corosync is restarted. if you have done change ip or...
  • UdoB
    UdoB reacted to gurubert's post in the thread Support for clustering using wireguard with Like Like.
    Keep in mind that clustering only works with latencies below 5 milliseconds.
  • UdoB
    Sure. Basically I am with @Johannes S : it is up to you to choose "the best" option. I try to go with ZFS if anyhow possible. To speed up the HDD pool a "Special Device" makes a really large difference - not only "measurable" but "feel-able"...
  • UdoB
    Well this depends heavily on your usecase and which comprimises you are willing to take in terms of performance/costs etc. I personally prefer to seperate between VM data (operating systems and applciation (including configuration e.G. for...
  • UdoB
    That version is very old and out of support for some time. Maybe test your hardware components and install a fresh PVE 8.4 and restore VM/CTs from backup?
  • UdoB
    In your first screenshot you have "iface eno1..." and "bridge-ports enp3s0...". While I am not sure if that is the reason for your problems it is a least inconsistent; try to correct it...
  • UdoB
    UdoB reacted to Johannes S's post in the thread Shared storage with Like Like.
    Even with a four nodes cluster it's recommended to have an external device for quorum (can eben be a small PC or a raspberry). For now you could just setup a small VM in your VMWare cluster for it, this would be more than sufficient: Another...
  • UdoB
    UdoB reacted to chrcoluk's post in the thread Disk cache none...safe or not safe? with Like Like.
    So from the link the OP posted. DirectSync is the safest, but expect trash performance. No host side caching, all writes sent to to disk force flushed, even if they are async. The None option is the default and most sane option, if using ZFS...
  • UdoB
    UdoB reacted to leesteken's post in the thread Disk cache none...safe or not safe? with Like Like.
    Slight nitpick: writeback should honor flushes (like fsync) and only buffer async writes. That's what differentiates it with writeback(unsafe) which fakes flushes and also buffers sync writes.
  • UdoB
    UdoB reacted to readyspace's post in the thread Disk cache none...safe or not safe? with Like Like.
    "None" (or "No cache") is generally considered safer than "Writeback" because: With cache=none, the guest OS communicates directly with the disk image, bypassing the host's page cache entirely. This reduces the chance of data being silently lost...
  • UdoB
    UdoB reacted to 6adminIT's post in the thread Commandline to show status of last backup with Like Like.
    Hello, thank you LnxBil for the script I made this one for nagios/centreon plugin from your work: #!/usr/bin/env python3 ######################################################################## ## ## Written by 6adminIT ## Based on LnxBil...
  • UdoB
    ??? What do you expect to get as the result? A fully working PVE is simply impossible. The lower core of PVE is the kernel. All containers share the same kernel, the one the host has loaded. A PVE inside of a container can not load its own...
  • UdoB
    We are pleased to announce the first beta release of Proxmox Backup Server 4.0! The 4.x family is based on the great Debian 13 "Trixie" and comes with a 6.14.8 kernel and OpenZFS 2.3.3. Note: The current release of Proxmox Backup Server 4.0 is a...
  • UdoB
    There's no such option because you usually don't want to use file based virtual disks. They are slow and CTs don't support snapshots with them. If you want more control you need to install it on top of debian. What's your issue with LVM? If it's...