Having previously restarted neutron-dhcp-agent on the controller and
watching my test instance continue to fail to get DHCP response, I can
now no longer induce this failure mode.  No difference in logs before
or after, no particular difference in load either.  ovs-ofctl
dump-flows on br-eth1 was also unchanged, I had suspected something
was going wrong there since I was seeing the requests on br-eth1-phy
but not on eth1 on that switch, but apparently that wasn't it...

Anyone have a clue what's different with network plumbing on instance
reboot from creation/rebuild?

-Jon

_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : [email protected]
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

Reply via email to