Created attachment 666964 [details] 0001-Add-support-for-building-RHEL-5-U9-images.patch Description of problem: Add support for building RHEL-5.9 instances to oz Additional info: * See attached "patch"
This is brewed as oz-0.8.0-7.
FYI, we discovered a issue today where instead of U9 for the os version, the user must specify 9. Spoke with ian and he is working on a respin. Current versions are at imgfac_1.0.2_3_gadcc5de-1 and oz-0.8.0-8
(In reply to comment #3) > FYI, we discovered a issue today where instead of U9 for the os version, the > user must specify 9. Spoke with ian and he is working on a respin. > > Current versions are at imgfac_1.0.2_3_gadcc5de-1 and oz-0.8.0-8 On a system with the following components installed ... * imagefactory-1.0.2_3_gadcc5de-1.el6cf.noarch * oz-0.8.0-8.el6cf.noarch I'm seeing that oz expects the <version>U9</version> field. > def get_class(tdl, config, auto, output_disk=None): > """ > Factory method for RHEL-5 installs. > """ > if tdl.update in ["GOLD", "U1", "U2", "U3"]: > return RHEL5Guest(tdl, config, auto, "rtl8139", None, output_disk) > if tdl.update in ["U4", "U5", "U6", "U7", "U8", "U9"]: > return RHEL5Guest(tdl, config, auto, "virtio", "virtio", output_disk) This appears expected to me. Is the problem elsewhere?
The problem was actually only regarding EC2 images, it existed in the ami id mapping config file specifying 9 instead of U9. /etc/imagefactory/jeos_images/ec2_rhel_jeos.conf good 2 go with [root@cloud-qe-14 ~]# rpm -qa | grep imagefac imagefactory-jeosconf-ec2-fedora-1.0.3-1.el6cf.noarch imagefactory-jeosconf-ec2-rhel-1.0.3-1.el6cf.noarch imagefactory-1.0.3-1.el6cf.noarch