Unable to move userdata to LXC mountpoint

[jwgn]

Hi,

there is no such option to move the userdata directory in dietpi-drivemanager.

Dietpi running on Proxmox in an LXC container. /mnt/storage is a mountpoint on the host, not a network drive.

What am I missing?

[MichaIng]

Hmm yeah this mount source, not being a /dev node, is interpreted as network device 🤔. How do you mount that drive in the first place? hdd/subvol-156-disk-0 looks like Proxmox volume, matches the naming schemes we discussed in context of our Proxmox VM installer: https://dietpi.com/forum/t/15679/50?u=michaing
But I have never seen how this is mounted. Is this done via Proxmox GUI only, per VM resp. container? How does the related /etc/fstab entry look like?

In any case, you can work around this for now:

/boot/dietpi/func/dietpi-set_userdata /mnt/dietpi_userdata /mnt/storage

[jwgn]

The command throws an error.

root@postgres:~# /boot/dietpi/func/dietpi-set_userdata /mnt/dietpi_userdata /mnt/storage/dietpi_userdata

 DietPi Updating user data location
─────────────────────────────────────────────────────
 Mode:

[ INFO ] DietPi-Set_userdata |  - From : /mnt/dietpi_userdata
[ INFO ] DietPi-Set_userdata |  - To   : /mnt/storage/dietpi_userdata
[ INFO ] DietPi-Set_userdata | Please wait...
[  OK  ] DietPi-Set_userdata | mkdir -p /mnt/storage/dietpi_userdata
[  OK  ] DietPi-Set_userdata | Free space check: path=/mnt/storage/dietpi_userdata | available=51189 MiB | required=11 MiB
[ SUB1 ] DietPi-Services > stop
[  OK  ] DietPi-Services | stop : cron
[  OK  ] DietPi-Services | stop : postgresql
[ INFO ] DietPi-Set_userdata | Moving your existing data from /mnt/dietpi_userdata to /mnt/storage/dietpi_userdata, please wait...
cp: cannot stat '/mnt/dietpi_userdata/./postgresql/15/main/postmaster.pid': No such file or directory
[ SUB1 ] DietPi-Services > start
[  OK  ] DietPi-Services | start : postgresql
[  OK  ] DietPi-Services | start : cron
[FAILED] DietPi-Set_userdata | Failed to copy /mnt/dietpi_userdata/ to /mnt/storage/dietpi_userdata.
[FAILED] DietPi-Set_userdata | Exited with error

The directory is a mountpoint created via Proxmox WebUI.

mp0: local-storage:subvol-156-disk-0,mp=/mnt/storage,backup=1,mountoptions=noatime,size=50G

It is explained here https://pve.proxmox.com/wiki/Linux_Container

Nothing in /etc/fstab:

# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0

root@postgres:~# mount | grep storage
hdd/subvol-156-disk-0 on /mnt/storage type zfs (rw,noatime,xattr,posixacl,casesensitive)

[MichaIng]

Okay that is some magic that is probably difficult to detect in any generic way. It does not appear in

lsblk
blkid

Right? Looks like best we could do is detecting it from the hdd/.*-[0-9]*-disk-[0-9] source scheme.

We really need to switch from cp to rsync in this script. Looks like PostgreSQL PID file disappears a little after the service was stopped (command query released). Please try to stop the database first, respectively just all services:

dietpi-services stop
sleep 2 # failsafe
/boot/dietpi/func/dietpi-set_userdata /mnt/dietpi_userdata /mnt/storage

[jwgn]

blkid shows nothing, lsblk shows the host disks:

root@postgres:~# lsblk
NAME     MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
sda        8:0    0   3.6T  0 disk
├─sda1     8:1    0   3.6T  0 part
└─sda9     8:9    0     8M  0 part
sdb        8:16   0   3.6T  0 disk
├─sdb1     8:17   0   3.6T  0 part
└─sdb9     8:25   0     8M  0 part
sdc        8:32   0 223.6G  0 disk
├─sdc1     8:33   0  1007K  0 part
├─sdc2     8:34   0     1G  0 part
└─sdc3     8:35   0 222.6G  0 part
sdd        8:48   0 223.6G  0 disk
├─sdd1     8:49   0  1007K  0 part
├─sdd2     8:50   0     1G  0 part
└─sdd3     8:51   0 222.6G  0 part
zd0      230:0    0    50G  0 disk
├─zd0p1  230:1    0   100M  0 part
├─zd0p2  230:2    0    16M  0 part
├─zd0p3  230:3    0  49.2G  0 part
└─zd0p4  230:4    0   696M  0 part
zd16     230:16   0     4M  0 disk
zd32     230:32   0    50G  0 disk
├─zd32p1 230:33   0   260M  0 part
├─zd32p2 230:34   0   512K  0 part
├─zd32p3 230:35   0     8G  0 part
└─zd32p4 230:36   0  41.7G  0 part
zd48     230:48   0     1M  0 disk
zd64     230:64   0    32G  0 disk
zd80     230:80   0    16G  0 disk

dietpi-services stop
sleep 2 # failsafe
/boot/dietpi/func/dietpi-set_userdata /mnt/dietpi_userdata /mnt/storage

That worked! Thanks.

root@postgres:~# dietpi-services stop

 DietPi-Services
─────────────────────────────────────────────────────
 Mode: stop

[  OK  ] DietPi-Services | stop : cron
[  OK  ] DietPi-Services | stop : postgresql

root@postgres:~# /boot/dietpi/func/dietpi-set_userdata /mnt/dietpi_userdata /mnt/storage/dietpi_userdata

 DietPi Updating user data location
─────────────────────────────────────────────────────
 Mode:

[ INFO ] DietPi-Set_userdata |  - From : /mnt/dietpi_userdata
[ INFO ] DietPi-Set_userdata |  - To   : /mnt/storage/dietpi_userdata
[ INFO ] DietPi-Set_userdata | Please wait...
[  OK  ] DietPi-Set_userdata | mkdir -p /mnt/storage/dietpi_userdata
[  OK  ] DietPi-Set_userdata | Free space check: path=/mnt/storage/dietpi_userdata | available=51189 MiB | required=11 MiB
[ SUB1 ] DietPi-Services > stop
[  OK  ] DietPi-Services | stop : cron
[  OK  ] DietPi-Services | stop : postgresql
[ INFO ] DietPi-Set_userdata | Moving your existing data from /mnt/dietpi_userdata to /mnt/storage/dietpi_userdata, please wait...
[ SUB1 ] DietPi-Services > start
[  OK  ] DietPi-Services | start : postgresql
[  OK  ] DietPi-Services | start : cron
[  OK  ] DietPi-Set_userdata | Successfully moved your data from /mnt/dietpi_userdata to /mnt/storage/dietpi_userdata.
[  OK  ] DietPi-Set_userdata | Completed

[MichaIng]

Great. PostgreSQL is probably an exception with a PID file in its data dir, usually they are blow /run. I'll think about whether to implement some failsafe sleep to the script, or switch to rsync directly, which handles files that disappear during execution better, AFAIK.

Regarding the Proxmox mount, do you know whether it is always prefixed with hdd/, or whether this can be different?

[jwgn]

No, it's just in my case hdd because thats how I named the pool.

root@ds01:~# zpool status
  pool: hdd
 state: ONLINE

So it starts with <poolname>/

[MichaIng]

Ah right, that was actually the same for VM storages. So looks like for container mounts it is then <name>/subvol-<id>-disk-<nr>. And for Btrfs subvolumes or "dir" storages is might be as well <name>:<id>/subvol-<id>-disk-<nr>.raw, with or without .raw ending.

EDIT: Ah not exactly the same. On VMs it was either <name>:vm-<id>-disk-<nr> (colon only, no slash) or <name>:<id>/vm-<id>-disk-<nr>.raw (colon between storage name and ID, followed by slash). Gonna be a bit generously about the possible elements of this naming scheme, or just testing through all storage types as mounts in an LXC container.