Exception during mounting need more than 0 values to unpack

asked 2019-05-31 20:01:12 +0300

This post is a wiki. Anyone with karma >75 is welcome to improve it.

Cinder stopped working with hyper-v

ERROR cinder.volume.drivers.remotefs [req-7b937913-5921-46fc-9bd5-750f0d2a4e51 - - - - -] Exception during mounting need more than 0 values to unpack

Openstack Newton Vanilla version We use MSI

Configuration

[DEFAULT]
my_ip = 10.0.20.169
auth_strategy=keystone
volume_name_template=volume-%s
#volume_group = cinder-volumes
glance_api_servers = http://10.0.20.56:9292
image_conversion_dir=C:\ImageConversionDir
verbose=true
debug=true
lock_path=C:\OpenStack\Lock\
transport_url=rabbit://openstack:password@10.0.20.56
log_dir=C:\OpenStack\Log\
log_file=cinder-volume-smb.log
enabled_backends = HYPERVMEX2
storage_availability_zone = HYPERVMEX2
default_availability_zone = HYPERVMEX2
allow_availability_zone_fallback = True

[database]
connection = mysql://cinder:password@10.0.20.56/cinder

[coordination]
backend_url = file:///C:/OpenStack/Lock/

#[key_manager]
#api_class = cinder.keymgr.conf_key_mgr.ConfKeyManager

[HYPERVMEX2]
volume_driver = cinder.volume.drivers.windows.smbfs.WindowsSmbfsDriver
volume_backend_name = HYPERVMEX2
smbfs_default_volume_format = vhd
smbfs_oversub_ratio = 1.0
smbfs_used_ratio = 1.0
smbfs_mount_point_base = C:\OpenStack\mnt\
smbfs_shares_config = C:\Program Files\Cloudbase Solutions\OpenStack\etc\cinder\smbfs_shares_list
#smbfs_pool_mappings = //cluster-share/HYPERV:HYPERVSF

[keystone_authtoken]
auth_uri=http://10.0.20.56:5000
auth_url=http://10.0.20.56:35357
memcached_servers=10.0.20.56:11211
auth_type=password
project_domain_name=Default
user_domain_name=Default
project_name=service
username=cinder
password=password


[cinder]
default_availability_zone = HYPERVMEX2
storage_availability_zone = HYPERVMEX2
default_volume_type = HYPERVMEX2

[nova]
region_name=RegionOne
auth_strategy=keystone
auth_type=password
auth_url=http://10.0.20.56:35357
project_name=service
username=nova
password=password
project_domain_name=Default
user_domain_name=Default


//clusterhfc.kcs.cloud/clusterhfc -o username=kcs\cinder,password=asdQWE123

2019-06-03 17:52:07.342 22240 INFO cinder.service [-] Starting cinder-volume node (version 9.0.1) 2019-06-03 17:52:07.375 22240 INFO cinder.coordination [-] Coordination backend started successfully. 2019-06-03 17:52:07.378 22240 INFO cinder.volume.manager [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] Starting volume driver WindowsSmbfsDriver (1.1.0) 2019-06-03 17:52:07.378 22240 DEBUG osloconcurrency.processutils [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] Running cmd (subprocess): qemu-img --help execute C:\Program Files\Cloudbase Solutions\OpenStack\Python27\lib\site-packages\osloconcurrency\processutils.py:349 2019-06-03 17:52:07.410 22240 DEBUG osloconcurrency.processutils [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] CMD "qemu-img --help" returned: 0 in 0.031s execute C:\Program Files\Cloudbase Solutions\OpenStack\Python27\lib\site-packages\osloconcurrency\processutils.py:379 2019-06-03 17:52:07.411 22240 DEBUG cinder.volume.drivers.remotefs [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] Loading shares from C:\Program Files\Cloudbase Solutions\OpenStack\etc\cinder\smbfsshareslist. loadsharesconfig C:\Program Files\Cloudbase Solutions\OpenStack\Python27\lib\site-packages\cinder\volume\drivers\remotefs.py:465 2019-06-03 17:52:07.413 22240 DEBUG cinder.volume.drivers.remotefs [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] shares loaded: {'//clusterhfc.kcs.cloud/clusterhfc': '-o username=kcs\cinder,password=asdQWE123'} _loadsharesconfig C:\Program Files\Cloudbase Solutions\OpenStack\Python27\lib\site-packages\cinder\volume\drivers\remotefs.py:497 2019-06-03 17:52:07.418 22240 ERROR cinder.volume.drivers.remotefs [req-117b7732-33ba-453d-8ced-11f858ee485c - - - - -] Exception during mounting need more than 0 values to unpack 2019-06-03 17:52:07.418 22240 ERROR cinder.volume.drivers.remotefs Traceback (most recent call last): 2019-06-03 17:52:07.418 22240 ERROR cinder.volume.drivers.remotefs File "C:\Program Files\Cloudbase Solutions\OpenStack\Python27\lib\site-packages\cinder\volume\drivers\remotefs.py", line 274, in _ensuresharesmounted 2019-06-03 17:52:07.418 22240 ERROR cinder.volume.drivers.remotefs self.ensuresharemounted(share) 2019-06-03 ... (more)

edit retag flag offensive close merge delete

add a comment

answered 2019-06-03 15:52:27 +0300

lpetrut
206 ●3 ●3

updated 2019-06-03 15:56:34 +0300

Hi,

That's quite odd. Could you please provide more details about the environment? Which OpenStack release are you using? Have you used the Cloudbase MSIs or installed the projects directly using the github repos?

May be worth also pasting your config (after removing any sensitive data, such as passwords, addresses, etc). You may also enable debug logging and provide some more log details (e.g. 5 minutes worth of log messages before the error occurs).

Thanks,

Lucian Petrut

edit flag offensive delete link

Comments

I updated the question with the information you asked for, thanks in advance

enavakn7 ( 2019-06-03 19:55:47 +0300 )edit

Ok, so debug logging is enabled and the driver seems properly configured. Can you please paste the log output that shows up before the actual error? It may give us some context information.

lpetrut ( 2019-06-03 20:47:36 +0300 )edit

I don't like asking this, but the easiest way to get straight to the error would be to make a minor change in the code here: https://github.com/openstack/cinder/blob/newton-eol/cinder/volume/drivers/remotefs.py#L277 By replacing LOG.error with LOG.exception, we'd get the full trace.

lpetrut ( 2019-06-03 20:54:05 +0300 )edit

I just pasted it

enavakn7 ( 2019-06-04 01:56:08 +0300 )edit

That's great, that helps us pinpoint the issue. Cinder fails to resolve the share host name (HYPERVMEX2). Can you double check that this name can be resolved (e.g. using nslookup)?

lpetrut ( 2019-06-04 09:18:13 +0300 )edit

To bring some context, eventlet embeds a copy of dnspython, which is used to resolve host names as opposed to using native primitives. If you're able to resolve the host name using nslookup, it's possible that we're dealing with a dnspython bug. You may consider upgrading eventlet to 0.22.

lpetrut ( 2019-06-04 09:21:00 +0300 )edit

I added more information. Thanks in advance

enavakn7 ( 2019-06-04 22:44:33 +0300 )edit

I got that wrong, sorry about that. The host name is clusterhfc.kcs.cloud. Did anything change after upgrading eventlet? For what is worth, you can explicitly request specific versions when installing packages with pip.

lpetrut ( 2019-06-05 09:53:28 +0300 )edit

There's a list of Openstack requirement versions that gets tested by the CIs: https://github.com/openstack/requirements/blob/stable/newton/upper-constraints.txt. We usually try to stick with it to avoid unexpected issues (e.g. regressions that may come with recent lib versions).

lpetrut ( 2019-06-05 09:55:52 +0300 )edit

see more comments

Exception during mounting need more than 0 values to unpack

1 answer

Comments

Your Answer

Question Tools

Stats

Related questions

Exception during mounting need more than 0 values to unpack edit

1 answer

Comments

Your Answer

Question Tools

Stats

Related questions

Exception during mounting need more than 0 values to unpack