Storage drive bricked!? Unable to mount a full disk
- Thread starter imaginebricking
- Start date
dir: local
path /var/lib/vz
content vztmpl,backup,iso
lvmthin: local-lvm
thinpool data
vgname pve
content rootdir,images
dir: storage
path /mnt/pve/storage
content iso,rootdir,vztmpl,snippets,images,backup
is_mountpoint 1
nodes pve
shared 0
Last edited:
Maybe @waltar got this right already (apologies if I am slow), but from what I can see (this is summary for myself or someone else if they take over), maybe also you:
You do not have a probllem that you (as described in OP) "cannot mount my physical storage disk which had previously been mounted".
Technically, you have your physically ~1TB (give or take, let's not worry about this now) drive mounted on the PVE node (host), it happens to be XFS (shown as sda1 on the host) it is indeed 100% full.
This is mounted on the host into
In there, you have your "virtual" disk for VM stored as QCOW2 image - this is basically a file on that XFS filesystem on that physical drive.
It has 1065314680832 bytes which is ~1TB and it appears to be completely filling up that entire host disk's filesystem of XFS.
You are then mounting this QCOW2 image into your VM (/dev/sdb on the guest as ext4 - that's filesystem of that image).
** I literally apologise if you knew all this, but it was not clear to me so far. **
At this point, I would probably take a step back, shut down the VM.
Then I would try to mount the QCOW2 directly on the host, you would need:
Let's see if that works out without errors.
You do not have a probllem that you (as described in OP) "cannot mount my physical storage disk which had previously been mounted".
Technically, you have your physically ~1TB (give or take, let's not worry about this now) drive mounted on the PVE node (host), it happens to be XFS (shown as sda1 on the host) it is indeed 100% full.
This is mounted on the host into
/mnt/pve/storage.In there, you have your "virtual" disk for VM stored as QCOW2 image - this is basically a file on that XFS filesystem on that physical drive.
It has 1065314680832 bytes which is ~1TB and it appears to be completely filling up that entire host disk's filesystem of XFS.
You are then mounting this QCOW2 image into your VM (/dev/sdb on the guest as ext4 - that's filesystem of that image).
** I literally apologise if you knew all this, but it was not clear to me so far. **
At this point, I would probably take a step back, shut down the VM.
Then I would try to mount the QCOW2 directly on the host, you would need:
Code:
modprobe nbd
qemu-nbd --connect=/dev/nbd0 /mnt/pve/storage/images/444/vm-444-disk-0.qcow2Let's see if that works out without errors.
Appreciate you for sticking with me this far. I'll have another look tomorrow when my brain is less fried. Thanks mate!
No worries, I am happy to follow up further too, just in case we are not around here at the same time, anyone can pick that up ... if only to copy out the valuable pieces (I suppose you do not have somewhere else to copy that file to give it more breathing space).
You should basically be able to have it then accessible as block device (on the host) at
/dev/ndb0, which you can then mount into a directory, e.g. mkdir /mnt/chest, then mount /dev/ndb0 /mnt/chest and see from there. 
Life got in the way, it wasn't "tomorrow" as I had hoped but I'm back to have another bash at this.
But when doing this:
ls -la of the /dev/ directory shows the following for the nbd0 entry
I've still got questions as to how I got into this situation in the first place. Is it simply the drive being full? Like I said, everything was working nicely before.
I can get to this point without errors
But when doing this:
I get:You should basically be able to have it then accessible as block device (on the host) at/dev/ndb0, which you can then mount into a directory, e.g.mkdir /mnt/chest, thenmount /dev/ndb0 /mnt/chestand see from there.
Code:
root@pve:~# mount /dev/ndb0 /mnt/chest
mount: /mnt/chest: special device /dev/ndb0 does not exist.
dmesg(1) may have more information after failed mount system call.ls -la of the /dev/ directory shows the following for the nbd0 entry
Code:
brw-rw---- 1 root disk 43, 0 Nov 4 20:28 nbd0I've still got questions as to how I got into this situation in the first place. Is it simply the drive being full? Like I said, everything was working nicely before.