Search results

https://forum.proxmox.com/threads/lxc-usb-passthrough-zwave-stick.30058/

I know this is an old Post but I have the same Problem - the storage in Question is a md raid with lvm on top. Deleting an disk image that was moved to another storage is not possible... Regards, Jonas /edit: Okay, one needs to remove the disk images from the vms ("Unused disk image")...

Hi, your load is under 12 and the CPU usage is mostly under 100%. I don't think a new Processor will speed it up. Maybe look if it's just one VM that's causing it. Regards, Jonas

You can see how much CPU the single VMs take. But as the CPU ist not really at 100% I don't think it will matter too much. Regards, Jonas

Yes you can't as in Yes, you can not share this to all nodes. Regards, Jonas

Add it to you /etc/fstab. Yes, you can't attach it to all nodes. Regards, Jonas

Hi, der maximale RAM ausbau bei der appliance sind 8GB - das wird schnell eng. Ausserdem scheint da nur eine MSATA SSD reinzupassen, kein weiterer Massenspeicher... Ansonsten ist sie halt langsam und sparsam - und auch nicht billiger als ein NUC oder ein HP Microserver. Für die geschilderten...

Das ist mittels mount.bind möglich. Ich würde das aber tendenziell eher über NFS o.ä. lösen. LG Jonas

Hi, sorry I wrote the wrong way around. Add the ceph storage to the last node, migrate the VM on that host to the ceph storage and then change host.. Regards, Jonas

If you want QCOW2 you need to add the storage as "Folder". Regards, Jonas

You can first try to change to host and then the storage... Regards, Jonas

...or put the VM disks on an NFS store... Regards, Jonas

Why not use NFS? Regards, Jonas

Ceph does not "stripe" but distributes all written Blocks (well technically RDB does that I guess but that's besides the point) to different Hard Drives. What performance do you get? What's your ethernet connection right now? Did you interconnect every node with 10g or is the 4th node connected...

As you will install PVE on a DOM: Install PVE on DOM. Don't do anything to the drives. After that you just go by the Documentation: >https://pve.proxmox.com/wiki/Ceph_Server Use 3x8 TB as OSDs (why not all 4?) and one SSD as cache: >https://pve.proxmox.com/wiki/Ceph_Server#Creating_Ceph_OSDs...

There are even more that tell you to not do it. If you know what you do - go for it. But it _will_ fail (my first failed after just 2 months..). USB disks are a no-go if you have a lot of write cycles - wich you will have with PVE. Regards, Jonas

ifconfig is not supported anymore: https://en.wikipedia.org/wiki/Ifconfig Some Linux distributions have deprecated the use of ifconfig and route in favor of the software suite iproute2,[3] which has been available since 1999 for Linux 2.2.[4] iproute2 includes support for all common functions...

There is a wiki entry for exactly this problem. Regards

Hi, ich denke fips will einfach eine SSD als OSD nutzen und fragt ob das einen Performancegewinn bringt. Nachdem bei Ceph jedes OSD performance bringt: Ja wird einen kleinen Performanceschub bringen, wie groß der ist kann ich aber nicht sagen. LG Jonas

.. und wenn die Frage nach der Downtime ist: So lange wie ein reboot auf dem server dauert... LG Jonas