Latest activity

  • F
    Is this method safe and efficient in the product environment?
  • jsterr
    /dev/mapper/mpathx are just dynamic aliases created by multipathd based on order of device discovery, they are not guaranteed to be consistent across hosts or even across reboots. Adding or deleting iSCSI targets will also cause them to be...
  • S
    Hello all, I have a homeserver running since the last 3 months and just recently, I attempted to perform gpu passthrough. Below specs may be relevant for the discussion: Motherboard : Asrock Rack x570d4u-2l2t (Virtualization enabled with IOMMU...
    • 1769349143785.png
    • 1769349261085.png
  • R
    Ah right, thank you for that. Since it's in the roadmap section in both cases we can assume that it is still in the pipeline and not officially released. I did look through both documents however I searched for "conntrack" and not "connection...
  • G
    Gilou reacted to LnxBil's post in the thread Thick Provisioning to Thin Provisioning with Like Like.
    This distinction is irrelevant for PVE or any other virtualization platform that does not use files for its storage, as @aaron already wrote. I would ALWAYS recommend to run trim/discard inside of the guest to really free up actually free space...
  • G
    Check cloud-init logs, also systemd units failures (other than the obvious network ones), cloud-init is quite fragile. Also, make sure you didn't block DHCP on the firewall, though if the request goes through, it's probably OK.. and check what...
  • B
    birdy reacted to Impact's post in the thread qemu guest agent install on ubuntu server with Like Like.
    That has nothing to do with the package not being found. apt update would have been of interest because it's definitely there.
  • L
    Here's what I would do: Rent a small virtual computer from Forpsi or WEDOS in the Czech republic for like €1-€2 Euros per month. 1vCPU and 1GB of RAM is enough for just Tailscale. Install Tailscale and set it up as one of your exit nodes. Then...
  • G
    It's been mentioned on PVE (not PDM: there), so make sure to have matching qemu and qemu-server versions, and up to date ones. I haven't tried production cross cluster migrations, but I saw that conntrack migration is on the changelog for 0.9...
  • D
    I don't think it was related to the update, it was more a problem started appearing at this time, but not caused by the update. I identified, that it seems I ran out of RAM on the host when the backup is running. I lowered the amount of RAM...
  • J
    Wenn ich mich richtig erinnere, hat mal ein Entwickler hier auf Nachfrage folgendes zum Thema rausgehen ( Reihenfolge größtes zu kleinstes IOPS): - Lokale SSDs, lokale HDDs, NFS-Netzspeicher, S3 Das setzt aber voraus, dass NFS und S3 auf der...
  • 6
    6equj5 reacted to news's post in the thread Auf verlorenem Posten ohne Netz with Like Like.
    Tipp lege dir bitte mal einen USB 3.0 -> NIC 1 GBit/s Netzwerkadapter zu. Meine sind noch aus der USB 2.0 Zeit, da konnte man 100 MBit/s erreichen. Das reicht auch aus, um ein System aus den Stand wieder ans lokale Netzwerk zu bekommen. Dann kann...
  • J
    Johannes S reacted to beisser's post in the thread Eingeschränkter Backup Upload Speed with Like Like.
    er verwendet allerdings nicht die storage box, sondern das object storage von hetzner (die storagebox kann meines wissens nach kein S3). evtl ist das hier die bremse.
  • J
    Johannes S reacted to UdoB's post in the thread Eingeschränkter Backup Upload Speed with Like Like.
    PBS braucht IOPS. Die Bandbreite ist fast egal, weil die Latenz alles kaputt macht. "Time per request: 1702951 microseconds." + "Uploaded 104 chunks in 177 seconds.". Also nullkommasechs IOPS... Disclaimer: ich habe das noch nie so aufgebaut.
  • J
    An interesting project, but... My main goal is to test configuration changes on a virtualized copy of the Proxmox cluster from my work. Therefore, I need the same Proxmox kernels, the same network topology, the same configuration files, etc...
  • J
    Note that the ZFS storage Thin provision option affects the refreservation property. Discard/fstrim will still work without it. I'm guessing for LVM-Thin it behaves similar to this...
  • J
    depends on the storage. If you import to ZFS for example and the "thin provision" option is enabled, then it will be thin, as in, zeros won't be written*. On RBD, you would need to do a trim/discard after the import, as, IIRC, on RBD, zeros will...
  • J
    This distinction is irrelevant for PVE or any other virtualization platform that does not use files for its storage, as @aaron already wrote. I would ALWAYS recommend to run trim/discard inside of the guest to really free up actually free space...
  • U
    uzumo reacted to LnxBil's post in the thread Thick Provisioning to Thin Provisioning with Like Like.
    This distinction is irrelevant for PVE or any other virtualization platform that does not use files for its storage, as @aaron already wrote. I would ALWAYS recommend to run trim/discard inside of the guest to really free up actually free space...
  • UdoB
    UdoB reacted to LnxBil's post in the thread Thick Provisioning to Thin Provisioning with Like Like.
    This distinction is irrelevant for PVE or any other virtualization platform that does not use files for its storage, as @aaron already wrote. I would ALWAYS recommend to run trim/discard inside of the guest to really free up actually free space...