UdoB's latest activity

  • UdoB
    UdoB reacted to fba's post in the thread VM recovery with Like Like.
    If you deleted the vm, the snapshot volumes are deleted too. Nothing to restore the vm from then.
  • UdoB
    UdoB reacted to Johannes S's post in the thread Notifications over SMTP with Like Like.
    sendmail nutzt selbst zum verschicken auch smtp. Ursprünglich war sendmail der erste Mailserver, warum alle späteren Mailserver (darunter auch postfix) noch einen Befehl "sendmail" bereitstellen, damit ältere Programme damit zurecht kommen...
  • UdoB
    UdoB reacted to Johannes S's post in the thread Notifications over SMTP with Like Like.
    Sofern man SMTP konfiguriert, werden standardmäßig die Systemmails von postfix über das SMTP-Notificiationtarget verschickt, ohne dass man groß an Postfix rumbasteln muss...
  • UdoB
    Isn't this what happens when cloning with consumer NVMe? kioxia kbg40zns256gb IO Pressure Stall 60% This doesn't happen with enterprise SAS. husmm3280ass200 IO Pressure Stall 1.4% Did you purchase an enterprise SSD? *The above shows the...
  • UdoB
    UdoB replied to the thread Notifications over SMTP.
    Ja, genau. Ich nehme einen meiner externen Emailprovider und lasse alle Email dort relayen. Anleitungen dazu gibt es zuhauf, eine konkrete kann ich - ohne selber zu suchen - aber nicht nennen. Postfix ist ein ausgewachsenes Mailsystem...
  • UdoB
    That's interesting as PVE uses an Ubuntu Kernel ;-) Possible workaround as a last resort: install Debian Trixie first, then install PVE - https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_13_Trixie - you would lose "OS on ZFS" though...
  • UdoB
    UdoB reacted to Falk R.'s post in the thread Notifications over SMTP with Like Like.
    Es ist simpel und schlank. Die Intelligenz hast du dann woanders.
  • UdoB
    UdoB replied to the thread Notifications over SMTP.
    sendmail gehört (per Default) zu Postfix: ~# dpkg -S $(which sendmail) postfix: /usr/sbin/sendmail Wie sendmail mit dem eigentlichen Postfix kommuniziert ist mir gerade unklar - und egal ;-) Wichtig ist, dass diese Konstruktion nur dann korrekt...
  • UdoB
    UdoB replied to the thread storage status unknown in cluster.
    You have pools with different names. (( And this prevents replication which is really annoying :-( )) Migration of running VMs allows me to select a specific storage on the target Migration of LXC does not offer this to me At the end...
  • UdoB
    UdoB replied to the thread Notifications over SMTP.
    Dein "allinkl" redet mit dem konfigurierten Emailserver, dazu hast du im Einrichtungsdialog deine Zugangsdaten eingegeben. Dieser Mechanismus funktioniert nur innerhalb des Proxmox-Kontextes. "sendmail" ist eine lokal laufendes Programm, welches...
  • UdoB
    UdoB reacted to Johannes S's post in the thread ProxMox VM show out of memory with Like Like.
    See also https://forum.proxmox.com/threads/fabc-why-is-proxmoxve-using-all-my-ram-and-why-cant-i-see-the-real-ram-usage-of-the-vms-in-the-dashboard.165943/
  • UdoB
    UdoB reacted to leesteken's post in the thread ProxMox VM show out of memory with Like Like.
    See https://pve.proxmox.com/wiki/Upgrade_from_8_to_9#VM_Memory_Consumption_Shown_is_Higher :
  • UdoB
    UdoB replied to the thread “HP EliteDesk 800 G4 Mini”.
    I do not know how many nodes you utilize. If you have only three of them: https://forum.proxmox.com/threads/proxmox-ceph-performance-with-consumer-grade-samsung-ssd.179948/post-834586
  • UdoB
    Supplementary: The prompt to do the emergency root login during boot is shown after another failure makes it unlikely the server will complete the boot process without administrator intervention. Logging in will give you a shell with enough...
  • UdoB
    Da ist schon der erste Fehler. RaidZ mit langsamen HDDs und Performant, das widerspricht sich komplett. Mit 4 HDDs im Mirror wirst du OK Performance haben, aber so eher Schneckentempo. OK Wenn du einen LVM Thin-Pool baust, ist das ein Block...
  • UdoB
    Good catch!! My brain stopped processing as soon as my eyes spotted the "870 QVO ssds", which happened before reading that 2x4tb per node o_O
  • UdoB
    That's worst design possible. Beside the too cheap devices... Look at one node: when one OSD fails, the other one on the same node has to take over the data from the dead one. It can not be sent to another node because there are already copies...
  • UdoB
    For once, AI is right :) Any consumer drive will have low Ceph performance due to rocksDB and sync writes, but those drives in particular are terrible for anything but PC archiving purposes due to it's small SLC cache and very slow QLC nand...
  • UdoB
    While it's true that 3-nodes is the bare minimum for Ceph, losing a node and depending on the other 2 to pick up the slack workload will make me nervous. For best practices, start with 5-nodes. With Ceph, more nodes/OSDs = more IOPS. As been...
  • UdoB
    Nice approach :-) Even though you have redundant switches I would recommend to prepare a separate wire (not a virtual LAN, 1 GBit/s is sufficient) for one of your multiple corosync rings. Disclaimer: your approach is multiple levels above my...