I’m a network admin helping out one of our desktop support guys out with some problems with multicast imaging. Our problem is very similar to this post:
https://forums.fogproject.org/topic/14969/multicast?_=1752946353612
whereas all of our workstations just hang on the partclone screen.
I have packet captures going at the client side and the server. I appears there is some cyclical unicast (seeing tcp handshake, then a reset, then repeat) https traffic between the client and server. there is no evidence of either side sending out a packet destined to the multicast rendezvous group address in this stalled state. However, a single packet is observed going to the rendezvous multicast group from the server sent upon the creation of the imaging task. But this happens at least a minute before our clients pxe boot and get to the partclone screen and it never repeats.
While the computers are in this state, both server and pxe clients are joined to the rendezvous group address. I have confirmed with icmp, and iperf2 traffic that packets destined to the group are received on all hosts in question (server and two pxe clients). so multicast routing is seemingly fine.
I’m hoping someone with experience with multicast imaging could create a connection flow chart to illustrate what traffic I should be seeing. A packet capture of a successful multicast imaging initiation would be fantastic too!
I’m also curious if there is a way to access a different tty while the partclone screen is open? The conventional way is not working.
Any help would be appreciated. I know you will ask for versions etc. I will get that from our desktop guy. Anything in addition to that you’d like to know about our setup, let me know and i’ll find out.
Thanks all!
