[HTCondor-users] Public and local IP address problem

[Ivan BorÅ <axivanax@xxxxxxxxx>]

I have an openstack instance running with ip addresses(command ec2metadata)
local-ipv4:192.xxx.xxx.x

public-ipv4:172.xx.xxx.x

To enable communication with the instance, I have to use public-ipv4 address.To allow condor talk with itself in the instance i need to use local-ipv4 address.So here is the problem, it seems the head node condor wants to connect to my instance through the local-ipv4 address(I think the problem is that condor uses local address for local management and broadcasts this address to the head node), which ofcourse is unreachable, how can I set that condor only access the outside instance through the public-ipv4 address.

CollectorLog:
MasterAd     : Updating ... "< workflow.dev.cloud >"
12/13/13 14:26:25 In OfflineCollectorPlugin::update ( 2 )
12/13/13 14:26:28 ScheddAd     : Updating ... "< workflow.dev.cloud , 192.xxx.xxx.x >"
12/13/13 14:26:28 In OfflineCollectorPlugin::update ( 1 )
12/13/13 14:26:28 SubmittorAd  : Updating ... "< ubuntu@xxxxxxxxxxxxxxxxxx workflow.dev.cloud , 192.xxx.xxx.x >"
12/13/13 14:26:28 In OfflineCollectorPlugin::update ( 11 )
12/13/13 14:26:31 StartdAd     : Updating ... "< workflow.dev.cloud , 192.xxx.xxx.x>"
12/13/13 14:26:31 StartdPvtAd  : Updating ... "<workflow.dev.cloud , 192.xxx.xxx.x >"

As you can see it uses the localipv4 address, though I need it to use the publicipv4.

Thanks in advance.