launch VM from image with volume creation on Equalogic

asked 2014-05-07 08:44:01 -0500

epsmed gravatar image

updated 2014-05-11 16:29:21 -0500

smaffulli gravatar image

Hallo

I set up a multi node openstack system. Controller 10.0.0.11, neutron 10.0.0.21, compute1 10.0.0.31, compute2 10.0.0.32, block1 10.0.041 I can create and work with new instances until I try to create or attach them volumes. I connect already (block1 node) a Dell Equalogic storage system, which is working well and I can create volumes.

As you see in the log file the problem is that after the correct volume creation the with the name: disk-8894b59c-d792-4724-9675-89dc64a3a621 the system tries to attach the volume:

  /dev/disk/by-path/ip-192.168.130.186:3260-iscsi-iqn.2001-05.com.equallogic:0-8a0906-1769c3806-55100000145536a3-disk-8894b59c-d792-4724-9675-89dc64a3a62-lun-0

to the new instance.

As you see one CHAR, the last one from the volume name is missing and of course the action fails. I cannot find any configuration error and believe that the Equallogic driver works properly.

Any idea why it happens ??

Thanks in advance

2014-05-07 15:10:41.353 6674 INFO cinder.volume.drivers.eqlx [req-bf833640-12ba-41b7-a303-7f8132334028 - - - - -] EQL-driver: executing "pool select default show"
2014-05-07 15:10:41.354 6674 INFO paramiko.transport [-] Secsh channel 6 opened.
2014-05-07 15:10:41.761 6674 INFO oslo.messaging._drivers.impl_qpid [req-bf833640-12ba-41b7-a303-7f8132334028 - - - - -] Connected to AMQP server on controller1.ipxgate.com:5672
2014-05-07 15:10:41.776 6674 INFO oslo.messaging._drivers.impl_qpid [req-bf833640-12ba-41b7-a303-7f8132334028 - - - - -] Connected to AMQP server on controller1.ipxgate.com:5672
2014-05-07 15:11:01.826 6674 INFO cinder.volume.manager [req-77c7554e-49b4-456b-9ef9-45b08eb2cc93 - - - - -] Updating volume status
2014-05-07 15:11:01.827 6674 INFO cinder.volume.drivers.eqlx [req-77c7554e-49b4-456b-9ef9-45b08eb2cc93 - - - - -] EQL-driver: executing "pool select default show"
2014-05-07 15:11:01.829 6674 INFO paramiko.transport [-] Secsh channel 7 opened.
2014-05-07 15:12:01.837 6674 INFO cinder.volume.manager [req-77c7554e-49b4-456b-9ef9-45b08eb2cc93 - - - - -] Updating volume status
2014-05-07 15:12:01.837 6674 INFO cinder.volume.drivers.eqlx [req-77c7554e-49b4-456b-9ef9-45b08eb2cc93 - - - - -] EQL-driver: executing "pool select default show"
2014-05-07 15:12:01.839 6674 INFO paramiko.transport [-] Secsh channel 8 opened.
2014-05-07 15:12:18.298 6668 INFO cinder.openstack.common.service [-] Caught SIGTERM, stopping children
2014-05-07 15:12:18.298 6668 INFO cinder.openstack.common.service [-] Waiting on 1 children to exit
2014-05-07 15:12:18.298 6674 INFO cinder.openstack.common.service [-] Caught SIGTERM, exiting
2014-05-07 15:12:18.302 6668 INFO cinder.openstack.common.service [-] Child 6674 exited with status 1
2014-05-07 15:12:18.771 6692 INFO cinder.openstack.common.service [-] Starting 1 workers
2014-05-07 15:12:18.773 6692 INFO cinder.openstack.common.service [-] Started child 6698
2014-05-07 15:12:18.776 6698 AUDIT cinder.service [-] Starting cinder-volume node (version 2014.1)
2014-05-07 15:12:18.777 6698 INFO cinder.volume.manager [req-71090868-6280-4e41-b022-9e9aeabf312b - - - - -] Starting volume driver DellEQLSanISCSIDriver (1.0.0)
2014-05-07 15:12:19.131 6698 INFO paramiko.transport [-] Connected (version 1.99, client OpenSSH_3.4)
2014-05-07 15:12:19.639 6698 INFO paramiko.transport [-] Authentication (password) successful!
2014-05-07 15:12:19.640 6698 INFO cinder.volume.drivers.eqlx [req-71090868-6280-4e41-b022-9e9aeabf312b - - - - -] EQL-driver: executing "cli-settings confirmation off"
2014-05-07 15:12:19.641 6698 INFO paramiko.transport [-] Secsh channel 1 opened ...
(more)
edit retag flag offensive close merge delete

Comments

What say logs /var/log/nova/compute.log on Compute Node and /var/log/nova/scheduler.log on Controller ? Format your output or upload somewhere text files , providing links. It's difficult to read your post now.

DanIzack gravatar imageDanIzack ( 2014-05-10 02:55:29 -0500 )edit

I have same issue for the this problem, any idea?

Billy Lam gravatar imageBilly Lam ( 2014-05-12 01:40:49 -0500 )edit

dropping a CHAR like that seems like a bug rather than a misconfiguration, have you tried reporting it as a bug? I've used the EQL driver for >1yr on Grizly/Havana/Icehouse & have not see this, only imediate difference is I'm using rabbitmq and you seem to use qpid (can't see why that would matter)

jproulx gravatar imagejproulx ( 2014-07-24 18:44:29 -0500 )edit

With EQL backend I created a volume using the same volume name 'disk-8894b59c-d792-4724-9675-89dc64a3a621' and attached it to a nova instance and I couldn't reproduce this problem. I noticed the disk path doesn't use the volume name like your logs instead it uses the volume id. Icehouse?

rajini-ram gravatar imagerajini-ram ( 2014-08-28 14:03:46 -0500 )edit