Search results

Sorry, there are a few sources of confusion around this, including different behavior between browsers between eg. firefox and chrome in the past. The 'origin' is meant only to "pin" the URL from which authentication should succeeds to a specific address. For a cluster, this doesn't make much...

This flag only disables the automatically (kernel-)generated link local address for the particular interface it is set on. It does not affect networking or the processing of IPv6 packets in any way. It's still up to the container to choose what to send over the interface. You'll need to disable...

This does sound like a 32 bit version of the 'stat' syscall being used in some place. Will need to investigate.

Can't give you any particular dates, but we will definitely upgrade. It'll mostly depend on how much work it is and how much it breaks ;-)

No, that's the init process, I meant the number in column 2 from ps aux | grep lxc-start | grep 101

Also, next time this happens, could you please look for the container's `lxc-start -F -n $VMID` process PID and look at the output of `ls -l /proc/$PID/fd`? There was an lxc package where `pct enter` would cause the start process to accumulate file descriptors, however, your version output shows...

This is probably because the raw lxc command line tools aren't really cgroupv2 aware. Will need to investigate.

It would be nice if you could share your config and how you configured it, so we could improve the error handling there to be more useful. As for how to fix it, that would be easier knowing when and where the error appears, and also, what your config looks like. The UI has an Auto-Fill button...

Generally lxcfs should get restarted, only reloaded. If you restart it (or start it manually like you did for debugging), all containers need to be restarted. Did you get any errors during the upgrade process? Do you see any journal messages when trying to start a container?

Multi-facet U2F verification had indeed a bug, verifying the appid vs the origin which of course cannot work in this case. As for webauthn with subdomains, this should work with the upcoming updates by setting the rp and id to the common tld, leaving out the origin (so each node fills it out for...

The -d single part here looks like a bad idea. This is basically the equivalent of a raid0 for data, you're only using raid1 only for metadata, so data simply cannot recovered (unless your VM internally has multiple copies and can do that by itself, for example if you're running databases...

Im moment werden subvolumes direkt ohne ext4 image nur dann verwendet, wenn man kein Größenlimit verwendet (wie beim normalen directory storage), da man sonst btrfs quotas benutzen muss, was bis noch nicht implementiert ist. Geht derzeit allerdings nur via CLI (0 als Größe angeben: pct create...

this is where, from the container's point of view, the data should show up

If you're new to these things I recommend using a VM for the time being, as with those you could use actual USB pass through via the GUI. Otherwise, it's roughly these steps: 1) setup /etc/fstab entry for the device with the `x-systemd.automount` option Eg. add something like: UUID="2587-CC93"...

As for your stroage, that should be storage (o↔r typo);-) (And for our API I do recommend using `1` and `0` for booleans)

Das sieht seltsam aus, was sagt # stat /dev/null

There's a difference between accessing a USB device or accessing its storage. The way you're passing the device through is the former, but it sounds like you want the latter. You should mount it on the host (via the file system's uuid to not also depend on the specific USB slot it's plugged in)...

^ openssh private key if you're sure it's actually the right file, convert to pkcs8: $ ssh-keygen -p -N '' -m pkcs8 -f /the/key (this will replace `/the/key`) but you should re-check how you created this

And what does the corresponding mpX line in the config look like?

The mknod feature flag only enables a set of harmless device nodes to be created. For tun you can only bindmount it, but it will not be possible to use `mknod` to create another and for now this is not directly supported as the `mknod` feature is currently not configurable any further.