yahoo-eng-team team mailing list archive

[Adam Gandelman <1065702@xxxxxxxxxxxxxxxxxx>]

** Changed in: cloud-archive
       Status: Confirmed => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1065702

Title:
  After folsom upgrade, instances can no longer access existing volumes.

Status in Cinder:
  Fix Released
Status in Cinder folsom series:
  Fix Released
Status in Ubuntu Cloud Archive:
  Fix Released
Status in OpenStack Compute (Nova):
  Invalid
Status in OpenStack Compute (nova) folsom series:
  Fix Released
Status in “cinder” package in Ubuntu:
  Fix Released
Status in “nova” package in Ubuntu:
  Fix Released
Status in “cinder” source package in Quantal:
  Fix Released
Status in “nova” source package in Quantal:
  Fix Released
Status in “cinder” source package in Raring:
  Fix Released
Status in “nova” source package in Raring:
  Fix Released

Bug description:
  After an upgrade from essex to folsom, attempting to attach volumes
  that existed pre-upgrade to instances does not work.  Attempting to
  attach an existing volume to an instance results in a traceback ending
  in:

  '.join(cmd))
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] ProcessExecutionError: Unexpected error while running command.
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] Command: sudo nova-rootwrap /etc/nova/rootwrap.conf iscsiadm -m node -T iqn.2010-10.org.openstack:volume-00000002 -p 
  192.168.20.9:3260 --rescan
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] Exit code: 255
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] Stdout: ''
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] Stderr: 'iscsiadm: No portal found.\n'
  2012-10-11 15:18:53 TRACE nova.compute.manager [instance: e1eb74b1-c063-48b5-ab3a-1ec935b26001] 

  It looks like compute is trying to find the volume with target named
  iqn.2010-10.org.openstack:volume-00000002.  However, the upgrade from
  Essex to Folsom migrates all volume IDs to UUIDs.  On the volume host:

   ~# tgt-admin --dump
  default-driver iscsi

  <target iqn.2010-10.org.openstack:volume-cd04dd3b-f8af-4400-bb91-8051aa05ef63>
  </target>

  <target iqn.2010-10.org.openstack:volume-9def01e3-0929-4735-932e-338a29374df3>
  </target>

  <target iqn.2010-10.org.openstack:volume-1eba5cc6-747b-4df9-810c-1f8debfd2dad>
  </target>

  The volumes table in the database, where I assume this connection information originates from on attach, still contains the old data:
  mysql> select provider_location from volumes;

  +-----------------------------------------------------------------+
  | provider_location                                               |
  +-----------------------------------------------------------------+
  | 192.168.20.9:3260,2 iqn.2010-10.org.openstack:volume-00000002 1 |
  | 192.168.20.9:3260,1 iqn.2010-10.org.openstack:volume-00000001 1 |
  | 192.168.20.9:3260,3 iqn.2010-10.org.openstack:volume-00000003 1 |
  +-----------------------------------------------------------------+
  3 rows in set (0.00 sec)

To manage notifications about this bug go to:
https://bugs.launchpad.net/cinder/+bug/1065702/+subscriptions