Description of problem: When trying to launch a VM or perform a 'nova image-list' an internal server error is produced (500). The trace from api.log is attached. Version-Release number of selected component (if applicable): rhel-x86_64-server-6-ost-3 How reproducible: Steps to Reproduce: 1. Fresh RHEL 6.4 install, updated. 2. puppet install of RHOSv3 using stackforge modules including the openstack module - controller.pp node file is attached. 3. Actual results: see attached trace dump Expected results: no error 500 Additional info:
Created attachment 802923 [details] stack trace from nova api.log
Created attachment 802925 [details] controller.pp puppet node file
How did you get these stackforge modules? Only those from packstack-modules-puppet package are supported AFAIK. Please provide details on howto reproduce this including obtaining the modules and installing them.
I had installed RHOSv2, Folsom, removed all the Folsom packages, then installed RHOSv3, Grizzly, and these errors were showing up. Only after I wiped the system clean - rekicked from scratch - did the errors go away. That's not very encouraging if anyone is every planning on trying to do an upgrade from RHOSv2 to RHOSv3. Something was getting left around from the Folsom install. The stackforge puppet modules we're using are from github...
OK, so this was an environment problem rather than client bug. I don't think this scenario (wiping Folsom, installing Grizzly) is very relevant, especially now with Havana out. Also, I think you should use RHOS puppet modules when reporting a bug against RHOS.
It is troubling that an 'yum remove <all openstack related rpms>' left bits around that messed up the glanceclient from Grizzly, but you're right, the Folsom->Grizzly upgrade path is less relevant now that we're getting farther along in time. Based on this, you can safely close this ticket as NOTABUG or whatever.