leutnantlurch

Well i learned my self that many cores doesn't mean that the machine is capable of handling the tasks you throw at them. The single core perfomance is also important.

Möp?

Hallo zusammen Ich hab' da mal ne Frage und zwar werden bei mir bei den verbundenen Netzlaufwerken die Speicherplatzbelegung nicht oder falsch angezeigt. Es betrifft hierbei ZFS Pools, jeweils im RaidZ-1. Mal nen Screenie was ich meine: Der Share "Software" liegt ganz normal im Array und wird auch relativ korrekt angezeigt, aber leutnantlurch, scan und ZoRoArt_media...ihr sehts ja selbst. So sehen die Pools in Unraid aus: scan und leutnantlurch liegen auf dem HDD-Pool, ZoRoArt_media liegt auf dem NVMe Pool. Hat jemand von euch eine Ahnung, woran das liegen könnte? Mein innerer Monk ist gar nicht befriedigt... Danke euch und Grüsse aus der Schweiz

After upgrading to x299 and a i9 10900X with 64GB of RAM (yes i need the pcie lanes ) these are the results with iperf: I would say, this is perfect! Thanks to @Vr2Io for the tipp, I will tweak my cpu. Case closed! 😃

What do you guys think about the i9-10900X? I would have the opportunity to buy a X299 Motherboard and this CPU + 64GB of Ram for a reasonable price.

Hey! Thanks a lot for your answer. More and more I have the feeling that my CPU wants to retire... I have currently observed that the urBackup backup 1.) takes a long time and 2.) drives the CPU to the limit in places, the Unraid WebUI does not respond and the other applications that I run on the server become very sluggish or crash (in the case of Authentik). I haven't had any major problems so far, but until recently I didn't have a 2.5 GbE connection either. However, since a stronger CPU would automatically mean a platform change, I would like to be sure that it is a "performance problem".

Hmm okay, I had a simmilar issue with the BliKVM, I fixed the problem with reformatting and reimaging the SD Card with the pikvm image. But unfortunately that doesn't help you

Do you have a BliKVM? Perhaps the PCIe version?

Yes, you're right, I need to provide more data here I have now tested with iperf3. CPU Governor "power safe" "on demand" "conservative" "performance" WIndows filetransfer on ZFS NVMe Pool (testfile generatet with fsutil) "power save" "on demand" "conservative" "performance" And the tests with the HDD ZFS Pool (only on demand and performance) "on demand" "performance" After the tests, I'm no longer sure whether the transfer speeds are really not satisfactory. Can you take a look at it and tell me whether this is complaining on a high level?

Hello everyone I have an interesting problem with Unraid and ZFS. Briefly about my setup, I have a Unifi network at home. My PC is connected with a 2.5 GbE NIC to a 2.5 GbE switch. My Unraid server is connected to the same switch with an Intel X520-DA2 via 10 GbE. I have a normal HDD array with single parity, a ZFS cache, a ZFS HDD pool and a brand new NVMe ZFS pool. The platform is an X99 mainboard with 32 GB DDR4 Ram and an Intel Xeon E5 2699 v3 CPU. Now to my problem, I have a rather modest transfer speed for all ZFS pools. My tests have shown that the transfer speed is dependent on my CPU generator. On "power saving" I get transfer speeds of about 145 Mbyte/s, on "on demand" I get speeds of about 180 Mbyte/s and on "performance" the speed increases to about 250-280 Mbyte/s from transfer start to transfer end. I find this particularly irritating with NVMe pools, as these SSDs are absolutely capable of a speed of around 280 Mbyte/s from the start. Now I wanted to ask whether the CPU is perhaps no longer capable of delivering such speeds? Thank you for the answers. Greetings from Switzerland!