Bug 1320414

I tested this on IPv6 enabled OSP8 and I got the same result. Attaching the volume.log

+++ This bug was initially created as a clone of Bug #1300853 +++

Description of problem:
Unable to attach volume to instance in IPv6 and SSL deployment

Version-Release number of selected component (if applicable):
openstack-tripleo-heat-templates-0.8.6-110.el7ost.noarch

How reproducible:
100%

Steps to Reproduce:
1. Deploy IPv6 and SSL enabled overcloud
2. Spawn an instance
3. Create a volume
4. Attach volume to intance

Actual results:
Volume doesn't get attached.

Expected results:
Volume gets attached.

Additional info:
Attaching the error log in cinder/volume.log

--- Additional comment from Sean Cohen on 2016-01-27 09:40:43 EST ---

(In reply to Marius Cornea from comment #0)
> Created attachment 1117016 [details]
> cinder.log
> 
> Description of problem:
> Unable to attach volume to instance in IPv6 and SSL deployment
> 
> Version-Release number of selected component (if applicable):
> openstack-tripleo-heat-templates-0.8.6-110.el7ost.noarch
> 
> How reproducible:
> 100%
> 
> Steps to Reproduce:
> 1. Deploy IPv6 and SSL enabled overcloud
> 2. Spawn an instance
> 3. Create a volume
> 4. Attach volume to intance
> 
> Actual results:
> Volume doesn't get attached.
> 
> Expected results:
> Volume gets attached.
> 
> Additional info:
> Attaching the error log in cinder/volume.log

I understand that this is LVM flow, and not Ceph one (that should be supported)
Since LVM is a reference implementation that should not be deployed in production I don't see this as a high priority for a zstream release.

Please clarify / add customer case.
Sean

--- Additional comment from Marius Cornea on 2016-01-27 09:53:22 EST ---

This issue was encountered during IPv6 testing. I think this will show up when deploying block storage nodes as well which is supported for production as far as I know.

--- Additional comment from Sean Cohen on 2016-01-27 11:40:09 EST ---

(In reply to Marius Cornea from comment #2)
> This issue was encountered during IPv6 testing. I think this will show up
> when deploying block storage nodes as well which is supported for production
> as far as I know.

Understood. 

See: https://access.redhat.com/articles/1535373#RHEL-OpenStack-Platform-Director

The default driver with Cinder is the internal LVM software iSCSI driver. This is the volume back-end which manages local volumes.

However, the Cinder iSCSI LVM driver has significant performance issues. In production environments, with high I/O activity, there are many potential issues which could affect performance or data integrity,

Red Hat strongly recommends using a certified Cinder plug-in provider for storage in a production environment. The software iSCSI LVM driver should be used and is only supported for single node evaluations and proof of concept environments. 

Adding Yogev to comment on the Ceph IPv6 testing (Ceph has IPv6 support since 1.2.3)

--- Additional comment from Sergey Gotliv on 2016-01-28 06:58:58 EST ---

Guys,

This issue is related to the LVM driver only! Unfortunately its really hard to say what exactly happened because Cinder failed to print an error message due to Unicode error, but most probably we failed to instantiate rtslib.NetworkPortal with IPv6 address. Some put that comment [1] in the rtstool.py for the good reason.

Marius,

Can I get an access to that environment to check my suspect?

[1]  # TODO(emh): Binding to IPv6 fails sometimes -- let pass for now.

--- Additional comment from Marius Cornea on 2016-01-28 08:21:44 EST ---

Sure, the undercloud can be accessed by:
ssh stack.44.97 -p 2900
credentials file is overcloudrc

then jump on the controller by:
ssh heat-admin.2.23

--- Additional comment from Lon Hohberger on 2016-02-04 09:55:25 EST ---

Not a 7.0.4 blocker - sgotliv notes that LVM is PoC