New Question
0

Cinder Newton

asked 2017-01-09 15:34:34 +0200

Matthieu gravatar image

Hello,

We have migrated our Windows 2016 Cluster from Mitaka to Newton. Everything is fine now, in one of our 2 region, the Cinder service is not working (that's fine with the same configuration in the other region). At the start up, the service is starting correctly but that's not possible to create a new volume. In the log the only thing is that it writes "Exception during mounting [Errno 11001] No address associated with hostname". The UNC path is working with the account used to mount it. We have also seen that the mnt folder is not created.

Matthieu

edit retag flag offensive close merge delete

Comments

I still have no explanation for this issue... I tryied to replace the server name by its IP on a local share (not a SOFS one), but I always have the same. Our other cluster in an other region is worknig fine with the same conficuration. Tryied to un-install and reinstall but was still failling.

Matthieu gravatar imageMatthieu ( 2017-01-30 18:57:13 +0200 )edit

2 answers

Sort by » oldest newest most voted
0

answered 2017-02-28 16:11:33 +0200

lpetrut gravatar image

Hi,

We had some issues as well because of eventlet monkey patching the getaddrinfo function, using dnspython as a replacement, which has some bugs (e.g. not parsing the host dns settings correctly).

Depending on your eventlet version, you should be able to avoid having it replace the getaddrinfo function. You can simply set the following env variable:

EVENTLETNOGREENDNS = yes

(https://github.com/eventlet/eventlet/...)

This will most probably be included in the MSIs as well.

Regards, Lucian

edit flag offensive delete link more
0

answered 2017-02-27 15:34:49 +0200

Matthieu gravatar image

I am not sure we can consider that as an answer, but I get my problem resolved.

I have revert back the getaddrinfo function in C:\Program Files\Cloudbase Solutions\OpenStack\Cinder\Python27\lib\site-packages\eventlet\support\greendns.py to the Python default. It seems there is a bug there. This function is which working on one of my cluster and not the other.

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower

Stats

Asked: 2017-01-09 15:34:34 +0200

Seen: 1,416 times

Last updated: Feb 28 '17