Recent content by pabernethy

To get your cluster started you likely want a dedicated network between the nodes. It needs no management whatsoever, so no pfSense, DHCP, DNS or anything. It will serve as your corosync network. Corosync is rather sensitive to latency, so you want to keep it as low as possible. And with no...

Use the same keyboard language, too.

Please try using the same locales on host, guest and client. There's quite some guesswork involved in NoVNC and using the same locales for all three parties improves the chances of it making correct guesses.

NoVNC or SPICE? What locales are set on the host, guest and client?

Be warned that no_root_squash is not without risk

The user you mount the NFS share as is not allowed to write to either the mountpoint or the share. Assuming you mount as root you may want to look into mounting as a different user, that you give write access on the share. Or configure a mapping on the server.

Stéphane Graber's blog has some instructions on this topic. Although I'm not quite sure what you hope to gain from it.

The PVEAuditor role is intended for that purpose. If you want to limit access to VMs and nodes (excluding storage, pools and access control) you can give the user access to only /vms and /nodes. For example: # pveum aclmod /vms --roles PVEAuditor --users audit@pve # pveum aclmod /nodes --roles...

If you want people to be able to help you need to give us more information. Exact error message, exact order of commands on all involved nodes. And the exact commands, of course.

If you have the rootfs on lvm-thin you don't need to do anything, except not use the space. With thin provisioning the size is a limit, not a fixed value. If you want to lower the limit you can do it manually: Make a backup! Verify that you can restore from the backup and the restored container...

As I already said, shrinking is not supported.

Are you trying to shrink the rootfs? Otherwise I'll need more info than "doesn't work."

Because someone might want to destroy the container and create a new one in its place, reusing the disks. E.g. if it has data that's still required but they want to switch the linux distro.

Are you using a container or a VM? pct is for containers, qm is for VMs.

That's why. Blindly guessing that you did a standard setup, local-lvm is the name of storage, not the name of the volume group. The VG is name 'pve' by default. So you'll want to use # lvremove pve/vm-102-disk-2 Don't force it. You can actually do this from the WebUI. Select the VM, select the...