thanksalternatively, you can do "yum -y upgrade" on all the old nodes, reboot and voila, all nodes are now up to the same versionnever tried rpmrebuild but i'd imagine that it would take a huge amount of time and resources on the compute host. if this is a lab environment, i suggest re deploying your set up again.right now, what you can do, is to extract/rebuild the rpms on your compute host and rebuild them to use in your internal static yum repo. seeandyyes you will need to stabilise the package versions in an openstack deployment. so you better make your own yum repo. but the horses have bolted already from the barn. so just keep it in mind for the future
http://unix.stackexchange.com/questions/140778/how-to-build-an-rpm-package-from-the-installed-filesOn Tue, Apr 14, 2015 at 8:06 AM, lei mei <meilei007@gmail.com> wrote:_______________________________________________Hi everyone,When I add a new compute node to the openstack which I deployed some months ago, I meet a problem about the package version incompatible. Detail thing is below:1. I prepare a clean centos 7 system and add the ip address to the packstack answer file.2. Run packstack3. Everything looks fine and I get the successful hint at last.4. But I find the nova-compute service can't start on new compute node with below log:nova compute service fail to start due to "Connection to the hypervisor is broken on host"
5. I checked the libvirt on compute node, find it has upgrade to the latest version but the old openstack use the old version. And a lot of packages on compute node have the newer version than the old openstack.So I want to know how do you add a new compute node to the old openstack avoid this package version incompatible issue? BTW, I use the default yum repo, so should I maintain a internal static repo for expand the openstack?-BRAndy
Rdo-list mailing list
Rdo-list@redhat.com
https://www.redhat.com/mailman/listinfo/rdo-list
To unsubscribe: rdo-list-unsubscribe@redhat.com
--
Link to me on LinkedIn