tripleo

Bug #1696866
Comment #4

Comment 4 for bug 1696866

Revision history for this message

Ben Nemec (bnemec) wrote on 2017-06-15:

I did some more digging yesterday but concluded that this is beyond my ability to debug in a reasonable timeframe. I no longer think os-brick is the problem though - it looks like the data passed in to os-brick is already bad, so I think it's something in Cinder or Nova that is not correctly handling the ipv6 iscsi address. I'm going to add both projects to this bug and let them determine who needs to fix what.

To recap, if we bracket iscsi_ip_address then oslo.config fails to validate the value, if we remove the brackets then cinder-volume starts correctly and is able to create the volume, but it can't be attached to the Nova instance. This appears to be because the iscsi port is being appended to the ipv6 address without first bracketing the address to separate the address and port portions (see the attached nova-compute logs).

Here are the RDO package versions I was testing with:

$ rpm -qa | egrep "nova|cinder"
puppet-cinder-11.1.0-0.20170605202131.257519d.el7.centos.noarch
openstack-nova-conductor-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
python2-cinderclient-2.0.1-0.20170320163530.d0790e3.el7.centos.noarch
puppet-nova-11.1.0-0.20170605232112.27baec7.el7.centos.noarch
python-cinder-11.0.0-0.20170521083557.cedf99f.el7.centos.noarch
openstack-nova-common-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-api-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-placement-api-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-cinder-11.0.0-0.20170521083557.cedf99f.el7.centos.noarch
python-nova-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-migration-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-scheduler-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-console-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
python2-novaclient-8.0.0-0.20170517113627.e1b9e76.el7.centos.noarch
openstack-nova-compute-16.0.0-0.20170521033533.99bd334.el7.centos.noarch
openstack-nova-novncproxy-16.0.0-0.20170521033533.99bd334.el7.centos.noarch

I did some more digging yesterday but concluded that this is beyond my ability to debug in a reasonable timeframe.  I no longer think os-brick is the problem though - it looks like the data passed in to os-brick is already bad, so I think it's something in Cinder or Nova that is not correctly handling the ipv6 iscsi address.  I'm going to add both projects to this bug and let them determine who needs to fix what.

To recap, if we bracket iscsi_ip_address then oslo.config fails to validate the value, if we remove the brackets then cinder-volume starts correctly and is able to create the volume, but it can't be attached to the Nova instance.  This appears to be because the iscsi port is being appended to the ipv6 address without first bracketing the address to separate the address and port portions (see the attached nova-compute logs).

Here are the RDO package versions I was testing with: