Red Hat Bugzilla – Bug 1276706
Issue getting an IP address on the cfme-5.5 appliance
Last modified: 2015-12-08 08:42:42 EST
There is a problem getting an IP address on the cfme-5.5 appliance (Only tested vSphere) in build 5.5.0.8. Workaround: To obtain IP address after booting up, please add the following line to /etc/sysconfig/network-scripts/ifcfg-eth0: DEVICE=eth0 Devel status 10/29: Devel made a fix in the kickstart that will ensure that we have the configuration we need on first boot and we just finished testing out both upstream and downstream appliances built with that fix in place. PR has been created and waiting to be merged: https://github.com/ManageIQ/manageiq-appliance-build/pull/72 Devel is still diagnosing the underlying issue that is causing us to end up with an invalid configuration after the build.
The root cause is the difference in tools writing the configuration files in question. /etc/sysconfig/network-scripts/ifcfg-eth0 is being written by the anaconda dracut module parse-kickstart file upstream. This means we write the device line we need here (https://git.fedorahosted.org/cgit/anaconda.git/tree/dracut/parse-kickstart?h=anaconda-19.31.122-1#n363) unconditionally. The same file downstream is being written by dracut's ifcfg module. We are running into issues with the "eth0" interface being recognized as a "kernel ethernet name" here (http://git.kernel.org/cgit/boot/dracut/dracut.git/tree/modules.d/45ifcfg/write-ifcfg.sh?h=RHEL-7#n109) and here (http://git.kernel.org/cgit/boot/dracut/dracut.git/tree/modules.d/40network/net-lib.sh?h=RHEL-7#n705) If dracut sees a "kernel ethernet name" it writes the HWADDR rather than DEVICE. One should be all we need for proper network configuration, but imagefactory removes the HWADDR here (https://github.com/redhat-imaging/imagefactory/blob/master/imgfac/FactoryUtils.py#L109) because we wouldn't want to specify a device MAC on a VM image (as the mac should be unique across multiple deployments) So the PR referenced above may be our best bet until we can either move to the new network interface naming scheme (http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/) which should remove the issue with dracue or alter our build process to do what we need it to do regarding network interface config.
Latest build still had issue. Moving back to ON_DEV
https://github.com/ManageIQ/manageiq-appliance-build/pull/76
New commit detected on ManageIQ/manageiq-appliance-build/master: https://github.com/ManageIQ/manageiq-appliance-build/commit/299b249d7ddb6ca0a6fd9be9c92c4f3332115e7d commit 299b249d7ddb6ca0a6fd9be9c92c4f3332115e7d Author: Nick Carboni <ncarboni@redhat.com> AuthorDate: Tue Nov 3 10:23:12 2015 -0500 Commit: Nick Carboni <ncarboni@redhat.com> CommitDate: Tue Nov 3 10:23:12 2015 -0500 Fix the line that conditionally prints DEVICE= The original PR used && in place of || https://bugzilla.redhat.com/show_bug.cgi?id=1276706 kickstarts/base.ks.erb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
This issue is no longer present on 5.5.0.9, the appliance gets an IP address "out of the box"
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2015:2551