KeyHelp Community

Hello,

I've checked all the internet and didn't found anything which might work in my case.

The thing is: I have Proxmox LXC environment which is very suitable for me. I use ZFS filesystem for all the stuff here, which is also extra suitable as I'm able to unify all my backups in one format.

However there is a problem with KeyHelp which I cannot bypass.
The problem is: it requires quota enabled which is not possible on LXC container when hosted on ZFS filesystem, as quotas are not supported with zfs subvolumes.

So, when installing there is an error, and when adding clients there is an error as well.
So my question is - is there any way to bypass that thing by disabling quota totally?

I don't really need it at all while it is mostly hosting some internal stuff, but watching the errors when adding users (which probably can by manually bypassed by changing status of user in keyhelp database) is nothing I need in my life, so... maybe I can just disable it and let it work with rest of functions.


Thanks
Lukasz

No you can not bypass or deactivate quota.
Quota is essential for KeyHelp.

You must edit your fstab in order to activate quota.
This board is full of examples. Just search for LXC.

I'd suggest another approach: Create a KVM VM, inside install a plain Debian with EXT4.

I have a similar setup at home, not with Proxmox, but with OMV, that uses the Proxmox kernel with ZFS.

One of the VM ist my testing and developing KeyHelp.

@Tobi setting quota for LXC is possible but it is impossible when ZFS is your file system on your base host. At least this is what Proxmox documentation say and discussion board of proxmox. So no matter what I do with fstab, I’m unable to make quota works when I’m basing my containers on ZFS. All the examples here are speaking only about LXC and yes, on LXC it can be done, but when your storage is on EXT4, not ZFS. I have tried all the things from this board but none worked, as well as every other solutions which normally will work on other filesystems.

@24unix this will ruin my whole backup solution I have now where I am just making ZFS snapshots for every container on server which works great and which is flexible and fast. But has some cons like this one. I’ve more considering putting kvm and setting zfs inside and backuping inside zfs partition but this will omit some zfs features I would like to stay with me.

theqkash wrote: ↑Mon 10. Apr 2023, 14:45 @24unix this will ruin my whole backup solution I have now where I am just making ZFS snapshots for every container on server which works great and which is flexible and fast. But has some cons like this one. I’ve more considering putting kvm and setting zfs inside and backuping inside zfs partition but this will omit some zfs features I would like to stay with me.

Why?
My KVM image reside on a ZFS Volume, so snapshots are possible without problems.

Yeah, it is possible, but snapshots of single kvm image isn’t that size-optimised like on LXC. On LXC I will have only increments between files so I can easily browse them and get single removed email from any snapshot in minutes. I’ll not be able to do it that easily on kvm and also kvm snapshots on zfs can be tricky because it can be bigger as it is working on single file.