Search results

I'd guess memory contention. Turn off ksmd or at least do a: echo 2 > /sys/kernel/mm/ksm/run

Depending on its write performance, that SSD might be okay for the Intent Log (more or less write cache). But you probably need to reset your expectations because, you're not going to see good performance with that system. Do some IOPS calculations and you'll see why. Especially look at write...

How much memory on the computer? For a read-mostly workload on spinning-rust drives, a good sized ARC is necessary.

It's easily found in the wiki and other documentation.

Do you have multi-queue enabled in the guest OS's network settings. (I'm guessing not.)

Give me a few months for next year's budget funds to be released. It works great on Epyc V1

You can do pve-zsync in a stand alone environment, just need to be able to SSH w/o a password if doing it to a different machine. And you can even do it to external storage. It's so much nicer than the QEMU/KVM backups.

Bummer. :/ File a bug through their actual bug reporter. That way more of their team will see it, and it can be better tracked.

I have a hack for my home lab to allow me to use pve-zsync for it, and in my production environment. ZFS snapshots are so faster and more efficient than the normal individual image backups. For home, I created a ZFS disk image on the NFS store and am using pve-zsync to it. It has worked fine...

"Lazy" umounting the filesystem usually does the trick. Try an "umount -l <path>" from the command line. (That's a lower-case L.)

Except I said "that's not enough for cluster ops and any sort of back end traffic" Which is more than cluster ops alone. *eyeroll* If the backend network link gets saturated, BAD THINGS HAPPEN TO THE CLUSTER. And a 100Mbps OVH VRack isn't quite the same as a set of servers on the same 100Mbps...

If you're using their vRack backend network, make sure it's not throttled to 100Mbps - that's not enough for cluster ops and any sort of back end traffic. iperf is your friend

Check out https://pve.proxmox.com/wiki/PVE-zsync

I went hunting, and it's this: https://pve.proxmox.com/wiki/Booting_a_ZFS_root_file_system_via_UEFI So, not ZFS size in particular, just how it plays with UEFI.

@wolfgang: Has there been a change? Previously there were problems booting from larger ZFS-mirrors

Consider a decent (small) SSD on each node for the journal. Otherwise your write performance will reaaally suck. As it is, it's not going to be great with only 6 spindles of spinning rust.

Did you install the firmware for it?

It's possible, and doable. Though a bit of a pain. :) Read up on PCI-passthrough. Lots of posts here and in the wiki. Good luck!

Bummer. What do you have set for the CPU hardware in the VM's settings? Make sure it matches your native system.

Might want to edit your post, or remove it if you can. Not a good model ... if you want to make a decent first impression here. :p