[SOLVED] How to remove Datastore (Directory)

R

Rotzbub

Member
Sep 18, 2020
11
6
8
31
Hi,
Proxmox
Backup Server 1.0-1

Put in a new Disk. Create Directory and add as Datastore.
  1. I can not find a way to remove a Datastore (in GUI)
    # proxmox-backup-manager datastore remove store1

    "Note: The above command removes only the datastore configuration. It does not delete any datafrom the underlying directory."
  2. The underlying directory is still available and I cannot find a way to delete this.
    even if I unmount/plug the disk it is still there.
How can I use the disk (again) to create a 'new' datastore ?
 
Hi,
Rotzbub said:
How can I use the disk (again) to create a 'new' datastore ?
Click to expand...

So you created a datastore on a disk and what to recreate one on the same, forgive me if I ask but why so?
After all, you'd get the same thing again.
 
t.lamprecht said:
Hi,


So you created a datastore on a disk and what to recreate one on the same, forgive me if I ask but why so?
After all, you'd get the same thing again.
Click to expand...
I created a Datastore via: Administration - Storage/Disks - Directory - Create: Directory - (checked "add as Datastore")

I cannot delete the Directory which is using the dev uuid and marking the disk 'in use'. eg:
  • want to rename the Directory
  • use another filesystem
  • decided to create a zfs (single disk)
  • upgrade to zfs (mirror)
  • ...
The disk is marked as in use and listed in Directory (Storage/Disks)
 
Ok. Some things like renaming can often done without recreating everything, that's why I asked.

If you want to clear the disk completely (delete all data on it again), you can zap the disk:

Bash: 
# WARNING: below is dangerous, wipes disk partitions, use with care!
sgdisk --zap-all /dev/sdX

Replace the /dev/ path with the respective device, triple check you got the right one and hit enter. After that, it should be useable again for creating a new storage+datastore with it.
 
Last edited:
  • Like
Reactions: carlosericel
Thanks for the quick reply! I can use the disk again.

Still:
  • how to rename the Directory path it is using
  • after sgdisk --zap-all /dev/sdX the Directory ist still listed in "Storage/Disks - Directory" as /mnt/datastore/<name>
 
  • Like
Reactions: Mevlock
Just re-checked and we have actually already support for that in the API, and there are some patches from a colleague open for review to add it to the web interface.

For now, as workaround please delete the systemd mount unit manually,for that execute:
systemctl disable mnt-datastore-<id>.mount

where the ID is the datastore ID of that mount you used, it needs to be escaped though, so either use CLI autocompletion to get it or the systemd-escape <my-id> command to get the escaped version of your ID.
 
  • Like
Reactions: Rotzbub
t.lamprecht said:
Just re-checked and we have actually already support for that in the API, and there are some patches from a colleague open for review to add it to the web interface.

For now, as workaround please delete the systemd mount unit manually,for that execute:
systemctl disable mnt-datastore-<id>.mount

where the ID is the datastore ID of that mount you used, it needs to be escaped though, so either use CLI autocompletion to get it or the systemd-escape <my-id> command to get the escaped version of your ID.
Click to expand...
Thank you. This solves my problem.
 
I use this lil script to remove a datastore located on a extern USB device.
Parameter is the Datastore Name

#!/bin/bash
error=$(proxmox-backup-manager datastore remove $1 2>&1 >/dev/tty)
if [ "$error" == "" ]
then
systemctl restart proxmox-backup proxmox-backup-proxy
fi
pfad=$(mount | awk "/$1/ {print "'$3'"}")
if [ "$pfad" != '' ]
then
error=$(zpool export $1 2>&1 >/dev/tty)
if [ "$error" == "" ]
then
rm -r $pfad
fi
fi

Important ! .. Dont run it from build in GUI Shell use SSH /Putty to run it.
 
mike2012 said:
systemctl restart proxmox-backup proxmox-backup-proxy
Click to expand...
That breaks/interrupts all current incoming backup jobs and more, always use systemctl try-reload-or-restart ...
But it shouldn't be required anyway...
 
Thx, in a prior version of PBS "zpool export" didnt work without.
Tested on 1.0-11 and works without

Script changed to

error=$(proxmox-backup-manager datastore remove $1 2>&1 >/dev/tty)
if [ "$error" == "" ]
then
pfad=$(mount | awk "/$1/ {print "'$3'"}")
if [ "$pfad" != '' ]
then
error=$(zpool export $1 2>&1 >/dev/tty)
if [ "$error" == "" ]
then
rm -r $pfad
fi
fi
fi

Script to attach i use

#!/bin/bash
txt=$(zpool import)
pool=$(echo "$txt" | awk '/pool:/ {print $2}')
state=$(echo "$txt" | awk '/state:/ {print $2}')
if [ "$state" == "ONLINE" ]
then
zpool import -f $pool
pfad=$(mount | awk "/$pool/ {print "'$3'"}")
if [ "$pfad" != "" ]
then
echo "
datastore: $pool
path $pfad" >> /etc/proxmox-backup/datastore.cfg
fi
else
echo $state
fi
 
So more than a year later and no way to delete an existing datastore from pbs?

Above comment:

where the ID is the datastore ID of that mount you used, it needs to be escaped though, so either use CLI autocompletion to get it or the systemd-escape <my-id> command to get the escaped version of your ID.
Click to expand...

didn't make much sense to me anyways... so any help would be appreciated..

With a GUI, you should think deleting, adding a datastore should be trivial.. Guess it still is not..

Also still can't get any nfs share to be used as a datastore which is odd in 2022.

>> anyway.. proxmox-backup-manager datastore remove <name> worked
 
Last edited:
I see it now. But I do feel tricked by the extremely odd and non-intuitive placement of the button. I guess this isn't meant for normal users.
On my screen, it's way off to the right where my eye would not normally go.
 

Attachments

  • Screenshot from 2022-10-13 13-41-53.png
    Screenshot from 2022-10-13 13-41-53.png
    87 KB · Views: 252
Last edited:
  • Like
Reactions: pmra
You must log in or register to reply here.
Top Bottom