Problem accessing VM conf files previous install

[KevinH]

Hi, recently I added and removed some disks from my homelab.

I had a samsung SSD I was frying to death so I replaced it with some old spinning disks.

I did an import on rpool, but since I did a new Proxmox setup, there was already a pool called rpool.
So I renamed it to prevrpool, overview:

root@homelab1:/mnt/prevroot/ROOT# zpool list
NAME        SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP  DEDUP    HEALTH  ALTROOT
data        464G   291K   464G        -         -     0%     0%  1.00x    ONLINE  -
prevrpool   464G  14.3G   450G        -         -     3%     3%  1.00x    ONLINE  -
rpool       294G   878M   293G        -         -     0%     0%  1.00x    ONLINE  -
root@homelab1:/mnt/prevroot/ROOT# zfs list
NAME                           USED  AVAIL     REFER  MOUNTPOINT
data                          33.0G   416G       24K  /data
data/vm-100-disk-0            33.0G   449G       12K  -
data/vm-101-disk-0              24K   416G       24K  /mnt/vm101
prevrpool                     14.3G   435G      104K  /mnt/prevroot
prevrpool/ROOT                3.70G   435G       96K  /mnt/prevroot/ROOT
prevrpool/ROOT/pve-1          3.70G   435G     3.70G  /mnt/root
prevrpool/data                10.6G   435G       96K  /mnt/prevroot/data
prevrpool/data/vm-100-disk-0  3.54G   435G     3.54G  -
prevrpool/data/vm-101-disk-0  7.04G   435G     7.04G  -
rpool                          878M   284G       24K  /rpool
rpool/ROOT                     876M   284G       24K  /rpool/ROOT
rpool/ROOT/pve-1               876M   284G      876M  /
root@homelab1:/mnt/prevroot/ROOT#

So I set the mountpoint, did an import, mounted pve-1 on /mnt/root

Next step would be to copy the configuration files of the VM's, but /mnt/root/etc/pve is empty.
Previously I always copied them while the system was running.
I see that on my new install it's actually a mount: /dev/fuse on /etc/pve type fuse,
which is in turn:
lrwxr-xr-x 1 root www-data 0 Jan 1 1970 qemu-server -> nodes/homelab1/qemu-server

but /mnt/root/etc/pve/nodes is not there either.

How would I proceed to get the configuration files?

I always thought this was just a plain directory, so I'm excited to learn why it's a fuse mount. What's the advantage? Some more info please??
Also, I've searched through the documentation, but maybe not in the right place. Is there something I missed relating to this issue?

 

[Stefan_R]

/etc/pve is a FUSE file system because it is managed by pmxcfs. This is used for cluster synchronization, but is also active when a node is running in standalone mode.

To your question on how to recover: I see two ways, the second one probably being the smarter one:

1. Copy the file /var/lib/pve-cluster/config.db from your old installation to your new one. This will probably not work straightaway, since the hostname and all your old storage settings are configured in that DB, so you'd have to edit it manually with some sqlite editor.
2. chroot into your old installation and run pmxcfs -l (local mode, no corosync or quorum required) and copy all files from /etc/pve/ to somewhere outside the FUSE mount where you can then copy them to their destination after leaving the chroot.