Assertion error after disk discovery

[captainfresh]

Im trying to run a test migration before committing to this solution but unable to get it to work - please see command and debug below.

root@ansible-ct:~/proxmove-main# proxmove cl-freshnet Freshnet nuc1 local-lvm pve-mail --no-verify-ssl --debug
2022-05-22 07:27:04,771: DEBUG: Parsing config file: ~/.proxmoverc
2022-05-22 07:27:04,772: DEBUG: (api) Connecting to dell-pve.freshnet.local
2022-05-22 07:27:05,139: DEBUG: (api) Connecting to nuc1.freshnet.local
2022-05-22 07:27:05,210: DEBUG: (api) 'Freshnet' nodes: [{'level': '', 'node': 'nuc1', 'disk': 4708163584, 'id': 'node/nuc1', 'maxdisk': 100861726720, 'ssl_fingerprint': xxxxxxxx, 'mem': 2295193600, 'uptime': 7378, 'type': 'node', 'maxcpu': 8, 'maxmem': 67322433536, 'cpu': 0.00581238172479048, 'status': 'online'}, {'uptime': 7238, 'type': 'node', 'maxcpu': 8, 'ssl_fingerprint': xxxxxxxx, 'mem': 2077593600, 'status': 'online', 'cpu': 0.00619684082624544, 'maxmem': 67322548224, 'level': '', 'id': 'node/nuc3', 'disk': 3895869440, 'node': 'nuc3', 'maxdisk': 62297829376}, {'maxdisk': 62297829376, 'id': 'node/nuc2', 'node': 'nuc2', 'disk': 3896066048, 'level': '', 'status': 'online', 'cpu': 0.00721536015653663, 'maxmem': 67322552320, 'maxcpu': 8, 'uptime': 7327, 'type': 'node', 'mem': 2116997120, 'ssl_fingerprint': 'xxxxxxxxx'}]
2022-05-22 07:27:05,210: DEBUG: Sanity checks and preparation
2022-05-22 07:27:05,210: DEBUG: Checking VMs existence on source and destination
2022-05-22 07:27:05,232: DEBUG: Checking for problematic config in 1 VMs to move
2022-05-22 07:27:05,238: DEBUG: (exec) ssh [email protected] /bin/true
[email protected]'s password:
2022-05-22 07:27:27,138: DEBUG: (exec) ssh [email protected] ls -l /dev/pve/vm-103-disk-0
[email protected]'s password:
Traceback (most recent call last):
  File "/usr/local/bin/proxmove", line 2374, in <module>
    main()
  File "/usr/local/bin/proxmove", line 2364, in main
    vmmover.prepare()
  File "/usr/local/bin/proxmove", line 1866, in prepare
    self.prepare_vm_config()
  File "/usr/local/bin/proxmove", line 1949, in prepare_vm_config
    self.prepare_vm_config_volumes(vm)
  File "/usr/local/bin/proxmove", line 1958, in prepare_vm_config_volumes
    elif volume.get_size('guest') is None:
  File "/usr/local/bin/proxmove", line 1688, in get_size
    assert size == size_property, (size, size_property)
AssertionError: (None, 8589934592)

~/.proxmoverc

[pve:cl-freshnet]
api=https://root@pam:[email protected]:8006

  [storage:cl-freshnet:local-lvm@dell-pve] ; local disk on node1 only
  [email protected]
  path=/dev/pve

; Example cluster named "the-new-cluster" with 2 storage devices; both
; storage devices exist on the respective nodes only.
[pve:Freshnet]
api=https://root@pam:[email protected]:8006

  [storage:Freshnet:local-lvm@nuc1]
  [email protected]
  path=/dev/pve

From the source the command
ssh [email protected] ls -l /dev/pve/vm-103-disk-0

produces...

lrwxrwxrwx 1 root root 8 May 14 17:37 /dev/pve/vm-103-disk-0 -> ../dm-16

Any help would be appreciated - thanks :)

[wdoekes]

Apparently you're ending up here:

proxmove/proxmove

Lines 913 to 927 in 6935ecf

	def get_volume_size(self, image_location):
	"""
	Read header bytes from qcow/qcow2 image to determine volume size.
	(The size observed by the VM guest.)
	"""
	if image_location.endswith('.raw'):
	# Raw image? Then its size is equal to the file size.
	return self.get_physical_size(image_location)
	elif image_location.endswith(('.qcow', '.qcow2')):
	# Extract the filesystem size from the qcow/qcow2 image below.
	pass
	else:
	# No support yet for other formats.
	return None

And then here:

proxmove/proxmove

Lines 895 to 909 in 6935ecf

	if image_location not in self._get_physical_size:
	path = os.path.join(self.path, image_location)
	# Use ls -l instead of stat because stat %s/%z is not
	# standardized across BSD/GNU.
	try:
	data = self.ssh_command(
	['ls', '-l', path], hide_stderr=True)
	except subprocess.CalledProcessError:
	log.warning(
	'Could not get image size from plain storage {!r}'.format(
	path))
	exact_size = None
	else:
	exact_size = int(data.split()[4].decode('ascii', 'replace'))
	self._get_physical_size[image_location] = exact_size

Which should be:

2022-05-22 07:27:27,138: DEBUG: (exec) ssh [email protected] ls -l /dev/pve/vm-103-disk-0

Which results in:

lrwxrwxrwx 1 root root 8 May 14 17:37 /dev/pve/vm-103-disk-0 -> ../dm-16

Which gets split and should become 8:

>>> data = b'lrwxrwxrwx 1 root root 8 May 14 17:37 /dev/pve/vm-103-disk-0 -> ../dm-16'
>>> int(data.split()[4].decode('ascii', 'replace'))
8

Which is useless, because we don't want the size of the symlink... but it shouldn't be None 🤔

For your dm-16, we'd probably need something like this instead of ls -l:

$ lsblk -bdno SIZE /dev/disk/by-id/dm-name-nvme0n1p3_crypt
511058116608

But, this should likely not be done by the ProxmoxStoragePlain class at all...

What does your vm config look like, and then specifically the configuration for the vm-103-disk-0 disk?

I cannot explain why:

• it would do an ls -l in the first place, as image_location.endswith('raw') should be False;
• so I'd expect a return None immediately (No support yet for other formats.);
• and when the ls -l is done, I would've expected 8 and not None; or a warning, which I don't see.

Did you change some code by any chance?

[captainfresh]

No change to the code at all (wouldn’t know what to do there 😝 ). I had to migrate the machine the other day by the backup restore method and its now sitting on a ceph cluster. The original machine was a standard issue Proxmox Mail iso install on local-lvm storage. Sorry I cant drag any more info up on that one.

[wdoekes]

Well. Then it looks like you're either in the need for this:
#33
or this:
#27
which are both marked as Wontfix, as I use neither.

I'm sorry. I don't have any quick fixes for you.

[captainfresh]

No worries, I appreciate you looking. -------- Original message --------From: Walter Doekes ***@***.***> Date: 25/5/22 20:08 (GMT+10:00) To: ossobv/proxmove ***@***.***> Cc: cfresh ***@***.***>, Author ***@***.***> Subject: Re: [ossobv/proxmove] Assertion error after disk discovery (Issue #44) Well. Then it looks like you're either in the need for this: #33 or this: #27 which are both marked as Wontfix, as I use neither. I'm sorry. I don't have any quick fixes for you. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>