Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1108362

Summary: Gateway Network Configuration setting is applied to the foreman host and /etc/dhcp/dhcpd.conf
Product: Red Hat OpenStack Reporter: James Slagle <jslagle>
Component: rubygem-staypuftAssignee: RHOS Maint <rhos-maint>
Status: CLOSED DUPLICATE QA Contact: Omri Hochman <ohochman>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 5.0 (RHEL 7)CC: mburns
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-06-11 19:59:43 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
output showing having to use the foreman hosts gateway
none
output showing using the gateway i want for the provisioned nodes, and the foreman host then losing all network connectivity none

Description James Slagle 2014-06-11 19:51:53 UTC
Description of problem:
When running staypuft-installer you're prompted for the Gateway during the initial Network Configuration.

This Gateway setting ends up getting applied to both the foreman host and 
I'm not sure if this Gateway refers to what the Gateway should be for the routers option in /etc/dhcp/dhcpd.conf.

While that may be correct in some cases, it most definitely is not correct in *all* cases.

It is likely that the gateway for the foreman host will be different than the gateway that will be to be set via dhcp for nodes going through the provisioning process.

In these scenarios, if you leave Gateway set to the correct gateway for the foreman host, you end up with the wrong gateway in /etc/dhcp/dhcpd.conf, which you will have to manually correct. I have attached output showing this in gateway1.txt. I'm able to complete the install, but the wrong value is in /etc/dhcp/dhcpd.conf, as I actually need option routers 192.0.2.1; there.

If you update Gateway to be the correct value that you need set in /etc/dhcp/dhcpd.conf during the staypuft-installer run, then that actually gets applied as the gateway to the Foreman host, and you lose all outbound network connectiving during the installation, causing packages to fail to download, and staypuft-installer will eventually fail. I've attached output showing this in gateway2.txt.

Version-Release number of selected component (if applicable):
[root@staypuft ~]# rpm -q foreman-installer-staypuft
foreman-installer-staypuft-0.0.15-1.el6ost.noarch


How reproducible:
I've reproduced both ways, numerous times, in different staypuft environments.

Here's my network setup:

I have the default libvirt network, 192.168.122.0/24 with dhcp, no changes from a default libvirt setup.

For the isolated provisioning network, I have a "brbm" libvirt network, which is defined from the following xml:
<network>
  <name>brbm</name>
  <forward mode='bridge'/>
  <bridge name='brbm'/>
  <virtualport type='openvswitch'/>
</network>

openvswitch is running on the libvirt host, and a bridge has been created called brbm:
sudo ovs-vsctl add-br brbm

The staypuft vm has 2 nics, eth0 on the libvirt default network and eth1 on the brbm network. Each vm that is to be provisioned by staypuft has 1 nic, and it's on the brbm network.

Comment 1 James Slagle 2014-06-11 19:53:05 UTC
Created attachment 907852 [details]
output showing having to use the foreman hosts gateway

Comment 2 James Slagle 2014-06-11 19:53:45 UTC
Created attachment 907853 [details]
output showing using the gateway i want for the provisioned nodes, and the foreman host then losing all network connectivity

Comment 3 James Slagle 2014-06-11 19:59:43 UTC
well, i spent a lot of time digging into this bug, but it appears there's already one filed. closing this one as a duplicate.

*** This bug has been marked as a duplicate of bug 1102706 ***