UdoB's latest activity

  • UdoB
    UdoB reacted to Johannes S's post in the thread PBS Backup to TrueNAS: How to do best? with Like Like.
    A zvolume (which is otherwise is the best option in terms of performance to go with a PBS VM on TrueNAS) doesn't profit that much from a special device as a dataset...
  • UdoB
    UdoB reacted to news's post in the thread PBS Backup to TrueNAS: How to do best? with Like Like.
    I expect 12*12TB of ZFS storage is only hdd based, then use also vdev zfs als special device. Example vdev0 zfs mirror 2x hdd, vdev1 zfs mirror 2x hdd and vdev2 zfs special device mirror 2x ssd. Example1: if you scale up your zfs dataset like...
  • UdoB
    I was under the impression that his TrueNAS exists already...
  • UdoB
    UdoB reacted to Johannes S's post in the thread Ext4 to zfs with Like Like.
    It would have been a better course of action to find out and solve the root cause. It's propably easier to backup everything and reinstall with ZFS on a single disc. But I would try again with a mirrored setup and trying to find the reason why...
  • UdoB
    UdoB reacted to aaron's post in the thread 5 Node Ceph Cluster without a switch with Like Like.
    We have customers who do run 5-node full-mesh clusters, for example with 4x 25Gbit NICs. Do not go for a ring topology as that could break in 2 places and then you have problems. The Routed with Fallback method is what you want...
  • UdoB
    Well..., you should be able to create a new zvol (not a "dataset" for files, but a block device) on the primary TrueNAS side. Then attach that device to the PBS VM. You have reached this goal when a new "sdb" (or equivalent) is visible inside the...
  • UdoB
    UdoB replied to the thread newbie questions about replication.
    "/etc/pve" is a fuse-mounted database. Proxmox does a lot of work to make sure that it is identical on all cluster members. Basically it reflects the current state of the cluster, including all configuration. That's also the reason why /etc/pve...
  • UdoB
    UdoB replied to the thread newbie questions about replication.
    That file (that VM with the ID 106) can only exist once in the cluster. To verify if the old one is still present: ~# find /etc/pve/nodes -name 106.conf - probably you'll find it there. Then move that one.
  • UdoB
    Sure. My friendly recommendation is to us a tool (here: PVE) the way it was meant to be used. Third party scripts are often problematic - both for the user (who does not know what it did in detail) and the people here in the forum because they...
  • UdoB
    UdoB reacted to 2000gtacoma's post in the thread Snapshots with Shared iscsi Storage with Like Like.
    All, This worked perfectly in my lab. I am now able to complete snapshots over iscsi with lvm. Thank you!
  • UdoB
    Klar, für die Konfigs ist ja auch der Cluster zuständig, solange der denn funktioniert. Aber ( wie du gerade gemerkt hast): Das ist kein Backup ( Cluster kann Probleme haben, wie jede Software), sondern um nach Ausfall eines Knotens die VMs und...
  • UdoB
    I am repeatedly astonished how beginners jump into the problematic areas immediately. Be it PCI-pass-through with GPUs for gaming or disks for TrueNAS or... third-party scripts. It seems that using PVE the-way-it's-meant-to-be is not sufficient...
  • UdoB
    UdoB reacted to LnxBil's post in the thread newbie questions about replication with Like Like.
    copy does not work on the filesystem behind /etc/pve, but a move will do. There is no solution for this besides adding more RAM to your second node to do this. If a single VM is to big to be started on the second node, you will not have fun with...
  • UdoB
    Iso-images are stored in a "Storage". The connection to any storages is defined in ".../pve/storage.cfg". The IP address (or DNS name) used in that configuration is used to connect to that server. But they are stored temporarily on a (PVE-)...
  • UdoB
    Das ist ein Kupfer-Adapter im M.2 Steckplatz, richtig? Migration belegt gerne so viel Bandbreite wie möglich --> "Sättigung". Das ist aber genau das, was man auf der Verbindung für Corosync auf gar keinen Fall haben will. Entschärfung: man kann...
  • UdoB
    Samsung 990 Pro 4TB. Szenario: Homelab, dass mittlerweile auch „etwas produktives“ hosted (aber nichts wo Geld oder Leben daran hängt) und per Tunnel und BGP ein öffentliches IPv4-Netz bekommt. Hardware: 2x Thinkcentre m715q Tiny Gen 2 und ein...
  • UdoB
    UdoB replied to the thread PBS WOL.
    Nur als Alternative zum WoL: viele BIOS-Varianten erlauben das Aufwecken gemäß einer eingestellten Uhrzeit. Herunterfahren kann der PBS dann entweder per stumpfer Crontab oder per Befehl vom PVE. Wie so oft: viele Wege führen irgendwo hin ;-)
  • UdoB
    Guten Morgen, ich lese hier nvme und ZFS ist ja schon mal eine tolle Sache, aber nur mit einem Datenträger? Gibt's doch kaum eine Möglichkeiten, wenn bit-rod stattfindet oder Flash-Zellen kaputt gehen, diese Daten wieder repariert zu bekommen...
  • UdoB
    Dann muss der Cluster schon früher auseinander gebrochen sein. Ich habe schon ganz viele Cluster in der Hand gehabt und keine Möglichkeit wie die Konfigurationen einfach verloren gehen, ohne das jemand von Hand etwas gemacht hat. Verrate uns was...
  • UdoB
    Das ist zwar grundsätzlich korrekt, aber auch bei einen single-node könnte einen die Konfiguration der VMs und Container flöten gehen, dagegen helfen halt nur Backups. Warum hast du Backup-Images der VMs und LXCs aber nicht von deren...