Search results

Indeed this is essential for a proper network connectivity. ;) Go ahead and good luck with your PVE project! :) Could you mark this thread as "solved"? Thanks!

This is cool and looks like that the storage systems aren't the bottleneck anymore but the network (1 GBit/s?).

I think this is as expected. Folder2Folder means that your HDD must read and write at the same time. This is a hard job for rotating platters. From NAS to VM your HDD is "just" writing.

ah, I get it - thanks! :)

Why do you want to change it? You're looking for IO performance, so raw is the better choice compared to qcow2.

And as Dominic suggested, I would choose "SCSI" instead of "VirtIO Block" as the vdisk bus/device type. As far as I know "VirtIO Block" is kind of deprecated. EDIT: And your CPU type is still "Default (kvm64)"? If you don't plan a cluster with different physical CPUs you should try "host".

kohle is the thread starter, not James Crook. ;)

Just to understand: What exactly is this SSD emulation? Is it always "better" to enable it, or what kind of tradeoff is to consider? Thanks for clarification and greets Stephan

How many of these disks are part of this RAID 1? Don't expect storage performance wonders from these disks, especially when it comes to "random small block" IO. Is this a "real world" file or 10 GB of zeros or something like that? Basically it's hard to benchmark such a complex environment...

slightly off topic: My first steps into the world of PVE clustering I did with a laptop installing PVE and three virtual PVEs inside. :D

Hey Reuven, what's the output of pvs, vgs and lvs? Greets Stephan

what the hell - thanks for sharing this experience! :oops: I knew that "bad hardware choice" can bring your performance down - but in such a way? Crazy...

almost :) How many and which disks are below this RAID 1? I guess, Dominic meant the bus/device type of the vdisk, not the controller type. As far as I know it's not possible to change this via WebGUI, so Dominic suggested to change it directly inside the configuration file of the VM. (Just out...

Hey kohle, what kind of storage (hardware) is running under "VMStore"? Greets Stephan

Of course this disk is far away from "the perfect PVE storage", but I don't think that this can explain the behavior described. Greets Stephan

Hey akus, some time ago I faced this problem - not with PVE but with KVM/Virt-Manager on a Linux desktop. Finally I discovered that I forgot to enable CPU virtualization features in BIOS. :rolleyes::D I'm not sure whether it's possible to install a PVE node without having CPU hardware...

:D Nur aus Neugierde: Sind die drei Server in Sachen Hardware identisch? Soll daraus ein Cluster werden? Dann rate ich Dir, falls möglich, die Netzwerk-Interfaces jeweils identisch zu konfigurieren, /etc/network/interfaces sollte also auf allen drei Kisten gleich aussehen.

Jep, genau so! In Deiner zweiten Ausgabe der Datei ist enp4s0f2 das Interface, das per DHCP mit einer Netzwerkkonfiguration versorgt wird, nicht mehr enp4s0f0 - Hast Du das (absichtlich) getauscht?

Ohne es ganz sicher zu wissen, behaupte ich frech: Nur die Interfaces, die in /etc/network/interfaces konfiguriert sind, sind von der Proxmox GUI aus sicht- und konfigurierbar. Ergänze also die Interfaces in /etc/network/interfaces und starte diese dann mit ifup <interface name>. Oder Du...

Und wie sieht /etc/network/interfaces aus? (Ich wette einen Keks, dass da auch nur acht statt zwölf Interfaces konfiguriert sind)