@gchartrandCRL For multicast to work properly all the hosts need to be “in sync”. Notice that when all the machines boot up they all wait on the blue partclone screen for everyone to join before it actually starts.

This is not only true for the first start but repeated on every partition. So all hosts need to wait and sync on the second, third, … partition. If a hosts arrives too late it can’t join this session and will drop out.

On the first partition FOG’s multicast manager waits for FOG_UDPCAST_MAXWAIT * 60 seconds but for all the other partitions the wait time is set to 10 seconds. This is usually enough but in some cases causes problems.

So to make a long story short, edit /var/www/html/fog/lib/service/multicasttask.class.php, jump to line 567 and change the value of 10 to 30 for example.

Before the change:

$i == 0 ? $maxwait * 60 : 10

After the change:

$i == 0 ? $maxwait * 60 : 30

Cancel all multicast tasks (or let them run to the end) and restart your whole FOG server to make sure it’s in a clean state. Then schedule a new multicast task and see if the same thing happens again.

@Tom-Elliott
i’ts working with the fog server 1.4.4 SVN 6077.

Thanks a lot.

@Tom-Elliott Thanks Tom, the update resolved the issue. I upgraded to 1.4.0-RC7.

@Tom-Elliott
No, no need an intermediate switch. If I turn-off spanning tree, yes I need to put a intermediate switch, but if it’s on, it’s working.

Finally, with all you explanation it’s really my image who’s corrupt. I was able to recreate it with another PC with the fog version 1.2 and now I can deploy it on my HP Prodesk G3.

Thanks for your time