Recent content by UdoB

Search for "fakeraid" in the search engine of your choice. Disclaimer: this is only a suspicion, I do not have an MS-01...

Try to first remove 6.2.16-20-pve to free up space in the ESP area - if you do not need this old kernel anymore. Possibly (not tested) by apt purge proxmox-kernel-6.2...

My personal opinion: YES, I want to get firmware updates and so I do enable that repo. "non-free" means it contains blobs without the source code being available. Debian hesitates (for a good reason) to include packages like this in the "main" or "security" repository. I am an update junkie...

Read: https://pve.proxmox.com/pve-docs/chapter-pmxcfs.html That filesystem is only writeable as long as Quorum is fulfilled. Did you establish a cluster? Possibly with two nodes? If yes: https://pve.proxmox.com/pve-docs/chapter-pvecm.html#_corosync_external_vote_support might help. Only if my...

My suspicion: this will also try to copy the folder "qemu-server" - and this fails. Notice that /etc/pve is not a folder but a fuse-mounted database. Not all semantics are implemented identically as for "normal" filesystems...

Please keep in mind: For two nodes to run in a stable setup you need a third node nevertheless. There are acceptable workarounds if you have only two "real" servers: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support Again: this is for the cluster in itself...

For two nodes to run in a stable setup you need a third node nevertheless. There are acceptable workarounds if you have only two "real" servers: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support Again: this is for the cluster in itself, not only for "HA"...

Not sure if this is your problem, but I believe backups may be smoother soon, as pve-no-subscription shows me: ~# apt changelog pve-qemu-kvm * implement support for backup fleecing :-)

We should differentiate between "backup/restore" and "live migration" backup/restore should work with any CPU, including "host", at least for most OS ;-) but live-migration (not asked for by @spleenftw) really needs the same CPU type as it copies register contents inside the CPU as part of the...

Try another stick without Ventoy. (Ventoy is really great, but I've also had strange problems with it in the past. Not sure which version, though.)

If you use sshfs to mount the old system's storage on the new PVE you save the need to supply additional (double-) space. You see... there's more than one way to skin a cat :-)

You are testing your RAM performance, your test does nothing storage-related: you are writing a lot of zeros, which are compressed to... a very few zeros befor they go to storage, some seconds in the future. For storage benchmarking search for fio. And no, ZFS on any kind of hardware raid is...

Please tell us what you did (so other users can learn it) and mark this thread "Solved" by editing the first post and selecting "Solved" in the drop-down menu for the title. Thank you.

Ja, CARP funktioniert erstaunlich gut, das war hier vielleicht ein eher unpassendes Beispiel. Ja, das Thema hatten wir ja schon öfter. Viel Ram und ECC und billig und wenig Stromverbrauch und viele Schnittstellen und 10GBit/s und ... existiert aber leider immer noch nicht. Choose your poison...

This is (probably) just buffer bloat. So live with it or shrink down (for test: just disable) your write caches. (No I have no idea how your storage has to be configured.) Best regards