Recent content by shalak

I believe once they add CT template for debian 13, they will also fix this perl script. Maybe their approach is "since we don't provide official trixie CT, we won't allow you to start one, if you dist-upgrade inside the container"?

Yup, decided to test this myself, and I can confirm: - created new CT 404 with additional non-rootfs mount - it automatically created /zbiornik-alpha/subvol-404-disk-0 - i started it, created some files - i stopped the container - run zfs rename zbiornik-alpha/subvol-404-disk-0...

Thanks! After zfs rename such dataset will no longer be managed by Proxmox GUI? I worry about situations like for example: I detach subvol-100-disk-0 from LXC 100 and then rename into it my-dataset, and sometime in the future, I'll remove LXC 100, then create a new one with same ID, it won't...

I don't get what you mean. I didn't post 2 methods, I posted 2 possible outcomes of the `mv` command and I'm wondering which is the correct one. But your clone suggestion got me thinking, what if I do a snapshot, clone it as shared-nas, then promote it and delete the subvol-100-disk-0? Will...

Hmm, not sure how this would help me. I understand, that I would need to - snapshot the subvol-100-disk-0 - clone it - mount the clone in e.g. /mnt/clone - and then mv /mnt/clone/* /zbiornik-alpha/shared-nas/ How does this change the situation? I'm still moving across datasets, am I not? It's...

Hi all, I have an LXC container on Proxmox that stores ~9.7TB of data under /zbiornik-alpha/subvol-100-disk-0/nas. This is a ZFS dataset with refquota=15T, and it's about 70% full. I’d like to move /nas into a separate ZFS dataset (/zbiornik-alpha/shared-nas) so it can be mounted into multiple...

and it doesn't say why is it not possible? again - it should tell you why. My bet is that you need to create the directory fist, i.e.: mkdir -p /etc/systemd/system/getty@.service.d

What is this "weird" reason? Maybe you don't have the /etc/systemd/system/getty@.service.d/ directory, you need to create it first. So you're modifying the main systemd file. It might (and will) be overridden by system upgrade. The "drop-in" files (i.e. files placed in the "*.d" directories)...

Huh, you effectively prevented yourself from local login, not sure if that's what you want to do. If you even lose SSH access, you'll probably need to restart into some single-user mode to fix this.

Well, if it prompts for the password it does not work :) You need to create a file that I mentioned above (/etc/systemd/system/getty@.service.d/autologin.conf) with the following content: [Service] ExecStart= ExecStart=-/sbin/agetty --noclear --autologin root %I $TERM You can add commands to...

I understand that option C makes backups to PBS extremely trivial, right? How much storage space would be lost to the ZFS overhead?

I understand, that in this scenario, the PBS can prone old ones? Ok, so basically, assuming that PBS is not compromised and it keeps last N snapshots, I don't need the PBS-in-the-middle to keep things secure. I guess that solves the issue. Thanks!

I believe that would defeat the purpose - the PVE in this scenario would need to have access to PBS host, thus exposing us at risk of encryption trojans Is it possible to have the following setup: - First PBS, running on PVE in a VM/LXC keeps only the latest backup - Second PBS, running on...

Here's my network overview: The VLAN 100 comes to physical eno1 port, trunk comes to eno2. Example VM is configured like this: Now, the problem is that I'm seeing constant network drops with periodic spikes. Here's is the view from the VM: The "drop" to 1.000 is a moment when I...

Indeed it did cross my mind, but that would require hooking up 2 ethernet cables, so I'll go with a smart socket. Thank you for the input :)