This resolves issue 12 Public IP isolation#2
This resolves issue 12 Public IP isolation#2dimm0 wants to merge 2 commits intotcooper:comet-modsfrom
Conversation
| returnxml.append(" <interface type='bridge'>") | ||
| returnxml.append(" <source bridge='%s'/>" % bridged_inter["device"]) | ||
| returnxml.append(" <mac address='%s'/>" % bridged_inter["mac"]) | ||
| returnxml.append(" <filterref filter='clean-traffic'>") |
There was a problem hiding this comment.
This looks too easy... Can you show us what did you do to test? Can you summarize how this is implemented in the host (i.e. dynamic iptables rules?). If iptables is restarted on the physical host what happens to this configuration? Do we need to add firewall rules to rocks and/or do an iptables-save on the physical host to prevent loss of this setting?
There was a problem hiding this comment.
Yes... please check. If I recall correctly this was one of the gotchas.
There was a problem hiding this comment.
I commented on the bottom in general discussion, otherwise this thread is shown in the code and it's not clear what's happening in the code.
tcooper
left a comment
There was a problem hiding this comment.
Please add testing / implementation details into the conversation.
|
http://libvirt.org/formatnwfilter.html#nwfexamples
Those other referenced filters are described in the same URL I first ran a VM without 'clean-traffic' parameter, and was able to change the IP inside the guest and access the network from it. Then I enabled the 'clean-traffic' parameter and was only able to access the network when the IP is set to the value defined by its "IP" parameter
I can check whether the rules stay when iptables are restarted |
|
|
|
Is there a reason to change network config or restart iptables while having VMs running on the host? Restarting NICs is a bad idea anyway when having passthrough enabled to the VMs... |
|
Perhaps... See sdsc/nucleus-service#12 |
Isolate public IPs for frontend VMs