Search results

If you are using the controller in HBA mode, you should be fine to use ZFS via the installer's advanced options [1]. [1]: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#advanced_zfs_options

We do not package Proxmox VE for ARM based ISAs. You seem to be using Pimox [1] which is a community project that is not officially supported by us [1]. [1]: https://github.com/pimox/pimox7

Hi, do you mean nothing gets restored at all order files that are downloaded have a weird name? Could you provide an example of such a file name? Also, I am assuming that you are only talking about file restore. So when you restore the entire VM, files are till in place, right? Could you...

Hi, this is a feature of your browser that hides the full path from our JavaScript components. This improves privacy and security [1,2]. The most we can realistically do is strip the “C:\fakepath\” part to only display the filename. You could open an enhancement request over in the bugzilla...

That looks correct, those permissions are actually 755, you missed an “r” there (you can check with `stat /etc/apt/trusted.gpg.d`). The issue remains strange, though. I'm guessing a curl --silent http://download.proxmox.com/debian/dists/bullseye/InRelease | gpgv --keyring...

Hey, yes you'd need two subscription keys if you want the whole cluster to be covered.

Well, usually you wouldn't populate two sockets with odd numbered CPU cores in real-life either. The socket option has two main uses-cases (from my knowledge at least): Helping pass through a more sensible NUMA topology to improve performance. This requires the NUMA option to be set as well...

This means that the number of sockets you configure for the VM should match the number of NUMA nodes of the host system the VM is running on. You can use `lscpu | grep "NUMA node(" to figure out how many NUMA nodes your host system has. Typically, the number of NUMA nodes of the host system...

Interesting, did you also configure the bookworm repos? What does `apt policy` report in terms of configured repos? Other than our and the regular Debian keys, is there anything else configured there that might be relevant?

Do you mean in your VM config? Usually not much, if your host system has two sockets, though, it may benefit from that if you also activate NUMA [1]. [1]: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_numa

Hey, I looked through the old thread and this looks strange. Are you sure the key is installed in the correct location and has the appropriate permissions? Could you check what `ls -lah etc/apt/trusted.gpg.d/proxmox-release-bullseye.gpg` and `apt-key list` output? It seems somehow apt struggles...

BTW: If you want an explainer of the differences between GPL and AGPL in more laymen's terms, the FSF has this article for you [1]. [1]: https://www.gnu.org/licenses/why-affero-gpl.html

This has already been reported and fixed [1]. The fix should be packaged with the next version of pve-manager. [1]: https://bugzilla.proxmox.com/show_bug.cgi?id=5734

So the question is, do you need a subscription for a QDevice? No, only for full nodes in a cluster. So if you have a two node cluster with a QDevice, only the two nodes need a subscription. The QDevice does not need an extra subscription.

What exactly are you worried about here? Looks fine to me :)

Nothing we provide officially, but generally CPUs of the same generation and vendor should work. Migrating from older to newer CPUs should also work in general. The other direction will likely not work unless you use a more restrictive CPU type. You can also create custom CPU types to improve...

Sorry, I can't provide anything definitive here, you might want to open a bug over at the Bugzilla [1], though. That's the preferred way of opening and tracking feature requests. [1]: https://bugzilla.proxmox.com/

In welchem Format erhalten Sie den die Daten von Azure? Je nach Storage kann das Zielformat auf Proxmox VE ein unterschiedliches sein. Auf file-based Storages können Sie entweder QCOW2 oder RAW verwenden. Wobei die beiden Formate unterschiedliche Features anbieten.

That won't work well. If you tell the VM it has 8 GB of memory, but actually only provide it with 4 GB it will likely crash or panic every time it's trying to use anything more than the 4 GB it actually has available.

Hi, we don't really support aarch64 officially, so this probably won't be fixed. However, that commit should only really be relevant for the backend. You can try to revert that commit if you just want to build the client as that change shouldn't be relevant for the client to function.