OpenStack status

Mark McLoughlin markmc at
Tue Nov 15 12:33:58 UTC 2011


And updates specifically from the past week or so:

   - We had some discussion about how to improve communication between 
     everyone working on OpenStack in Fedora and encourage more folks to
     help out:

   - An "Open Source Cloud and Virtualization" dev room has been 
     announced for FOSDEM in Brussels on Feb 4-5th:

   - A big openstack-nova update was pushed with ~20 significant 
     updates from the upstream stable branch:

   - Pádraig has continued the work on EPEL6 and posted an update:

       The current status is that I have nova and glance services
       running and passing very basic tests. I'm using 2 packages not
       yet in EPEL though.

         python-webob: 0.9.6 in RHEL6.2, 1.0.8 in F1[56]
         python-sqlalchemy: 0.5.5 in RHEL6.2, 0.6.8 in F15

   - Bob Kukura has been working on packaging Diablo's Quantum and will
     post a package for review soon. This has been complicated by the 
     fact that Quantum in Essex has undergone major changes to make it 
     easier to package.

   - Russell Bryant has been looking into how we might go about 
     running OpenStack with Apache Qpid rather than RabbitMQ.

   - Alan Pevec is preparing a keystone update based on the latest 
     upstream stable branch. He has also been debugging various issues 
     with keystone.

   - Cole Robinson has been working on getting Dashboard running on 
     Fedora and has also run into keystone issues.

   - I hit an issue with attaching a volume to an instance running on a 
     remote compute node:

     There's a fix proposed in gerrit which can be pulled into Fedora.

   - There's been a little bit more progress on making Nova use

     Rich and Pádraig are working up a plan. Also, Andrey from Grid 
     Dynamics pointed out that we actually really need this for EPEL6
     since the nbd kernel driver isn't supported on RHEL6.

   - Rich also posted an analysis of Nova's file injection code:

   - Derek Higgins got a fix merged upstream to have an instance's 
     "managed save" image removed when the VM is undefined:

   - The weird AMQP issue we saw once or twice during the test day has 
     undergone serious debugging by folks upstream, but the root cause 
     hasn't been identified yet:

   - Russell took a look at the "database is locked" errors we see with 
     sqlite and concluded that using SQLite is probably hopeless and we 
     should switch to MySQL by default:


More information about the cloud mailing list