Search results

Bei der PLP geht es aber darum, dass dann deine SSD Faktor 100 oder so schneller bei Sync Writes ist. Das hast du nur mit SSDs mit eingebauter PLP, nicht bei einer consumer SSD an einem UPS.

Best way to not waste RAM is to not assign a VM more RAM than it needs. 4GB should be enough for OMV unless you use ZFS or similar. Otherwise it will eat all RAM you throw at it which then isn't available for other VMs which might benefit more from it.

Hi, I needed to replace the hardware of a cluster node. The old node was using proxmox-boot-tool (systemd boot, no secure boot) and ZFS for the root filesystem. The "new" hardware only supports legacy boot. I would prefer to reuse the old disks without reinstalling PVE. Did someone already...

And again: https://www.linuxatemyram.com For Linux VMs, PVE webUI will show RAM used for caching as "used" and won't show "available" RAM as "free" RAM.

Some servers don't use proprietary hardware and will have standardized connectors and E-ATX/ATX/uATX formfactor (I like those older supermicro boards). Combine those Server mainboards with ECC RDIMMs, Server CPU, Enterprise SSDs and some quiet ATX PSU, tower cooler and case and you have a...

One problem is TrueNAS support. If you got a problem and ask in the TrueNAS forums they will roll their eyes and tell you that TrueNAS should be run bare metal and that the only supported hypervisor is ESXi. I personally got some problems in the past with shutting down the VM where it gets stuck...

Those values probably aren't temperature in degree celsius but the "health" for that metric going drom 0 to 100 or reverse. See the "smartctl -a ..." output and have a look at the datasheet of the disk manufacturer in case there is one that explains how to interpret SMART attributes.

TrueNAS in a VM works fine. But you probably want a dedicated HBA card and disks for TrueNAS for PCI passthrough so TrueNAS is able to directly access the real physical drives. TrueNAS uses ZFS and ZFS is quite demanding (lots of RAM, ECC RAM, Enterprise SSDs with PLP, CMR HDDs, lots of CPU...

Ok, because adding or removing any PCIe device might change the network interface names causing the network configuration to fail and then not allowing you to access the webUI.

NVMe SSDs or additional HBA/raid controller added?

Your pool only knows "nvme-eui.000000000000000100a0752446a4a08c-part3". So try something like: "zpool attach -f rpool nvme-eui.000000000000000100a0752446a4a08c-part3 /dev/disk/by-id/nvme-INTEL_SSDPE2KX010T8_BTLJ908508XJ1P0FGN-part3"

Point it to the actual device: sgdisk /dev/disk/by-id/nvme-Micron_7450_MTFDKCB800TFS_240446A4A08C ...

Like already said, ticking the "discard" checkbox alone won't help. You will have to configure your Ubuntu to trim free space. You will have to edit your Ubuntus fstab and mount filesystems with the discard option or create a systemd service/cron job to run a "fstrim -a". Using the latter it...

In the PVE webUI, go to the hardware tab of every VM, edit each virtual disk there and check the "Discard" checkbox. In addition you need to tell each guestOS inside your VM to actually trim/discard. How to do this completely depends on the guestOS your VM is using. In a Linux guestOS you could...

HDDs werden nie ansatzweise so schnell sein wie SSDs...egal was du da tust. Will man es performant haben ersetzt man die HDDs durch SSDs. Nicht vergessen, dass da eine schnelle HDD vielleicht 200 IOPS schafft und eine schnelle SSD 2,5 Mio IOPS. Selbst eine lahme SATA SSD ist da 400x schneller...

Correct Use "ls -la /dev/disk/by-id" instead of your "cd /dev/disk/by-id; ls". An SSD could have multiple IDs pointing to the same SSD. Yes. It will clone the partition table and both disks will have the same partitions. So you might want to install PVE to the smaller one so cloning won'T fail...

As I already said, its the same as replacing a failed disk of a mirror, you just don't replace a disk but add a new one. See the article linked above. So:

Ich würde den SLOG nie auf eine Platte mit Daten, die du nicht verlieren willst, packen. SLOG auf Systemdisk ist also eine ganz schlechte Idee. SLOG ist ja gerade das, was dir in Null-Komma-Nichts die SSDs killt und die Systemdisk möchte man ja im Normalfall nicht verlieren ;) Dann bitte was...

Jup, da muss man dann warten bis sowohl PVE als auch PBS auf die neue Major-Version aktualisiert wurden. Daher eher unschön. Auch kann man ein Bare Metal PBS nicht so leicht sichern. Aber LXC sind noch unschöner, weil PBS nie ausgelegt wurde um in einem LXC zu laufen und auch niemand vor dem...

You can't add a disk without formating it first and cloning the bootloader. And thats partition 3 of your physical disk because partition 1+2 are needed for the bootloader. See chapter "Replacing a failed bootable device" but instead of the "zpool replace..." you use a "zpool attach" for a...