Multicast remains on Partclone screen (HP switch)

[Tom Elliott]

@Matthieu-Jacquart I believe you’d only need to enable it at the thing handling the routing so it can move about. This namely because the main central switch or router is what’s communicating to the other switches. Of course enabling forwarding on each switch shouldn’t hurt either but I believe you’d need it pointing at the place to forward ad welk

[Tom Elliott]

@Matthieu-Jacquart assuming you have all vlans looking at a single fog server each vlan would need to allow forwarding to the broadcast address set for the fog server.

[Matthieu Jacquart]

@Tom-Elliott ok, appliying this command to all vlan (64158 is multicast port in fog configuration)

ip forward-protocol udp 192.168.10.60 64158

And now :

show ip forward-protocol
 IP Forwarder Addresses
    UDP Broadcast Forwarding: Enabled
 VLAN: 1
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 2
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 3
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 4
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 5
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 6
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 7
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 8
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 9
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 10
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158
 VLAN: 11
  IP Forward Addresses UDP Port
  -------------------- --------
  192.168.10.60        64158

But always same result, even with 1 PC
And I’ve just noticed that FOG_UDPCAST_STARTINGPORT seems change every time ???

[Tom Elliott]

@Matthieu-Jacquart it does, however you can override it in fog settings

[Matthieu Jacquart]

@Tom-Elliott you mean in file fogsettings on debian server, or in web management ?
Because I’ve tried to fix it in web management but it changes again…

[Tom Elliott]

@Matthieu-Jacquart in GUI. There’s actually a variable called port override

[Matthieu Jacquart]

@Tom-Elliott nop, no change…
Anyone else use fog with HP 2920/2530 swtich ?

[Matthieu Jacquart]

Ok, I’ll continue test with sotfware IC_Tester_Multicast and communication is ok between 2 PC (1 server and 1 client).

After that I run on fog server (with client in debug mode and “udp-receiver” command) :

root@FOG:~/fogproject/bin$ udp-sender --file /opt/fog/log/multicast.log --ttl 1
Udp-sender 20120424
Using full duplex mode
Using mcast address 232.168.10.60
UDP sender for /opt/fog/log/multicast.log at 192.168.10.60 on eth0
Broadcasting control to 192.168.10.255
New connection from 192.168.10.200  (#0) 00000009
Ready. Press any key to start sending data.
Starting transfer: 00000009
bytes=      17 782  re-xmits=0000000 (  0.0%) slice=0112 -   0
Transfer complete.
Disconnecting #0 (192.168.10.200)

So everything seems good for multicast, but still same problem : computer stays on partclone screen.
In multicast.log

[10-19-16 11:23:22 am]  | Task (16) Multi-Cast Task is new!
[10-19-16 11:23:22 am]  | Task (16) Multi-Cast Task has been cleaned.
[10-19-16 11:23:22 am]  | Task (16) /images/Win10 image file found.
[10-19-16 11:23:22 am]  | Task (16) 1 client(s) found or to receive.
[10-19-16 11:23:22 am]  | Task (16) Multi-Cast Task sending on base port: 64158
[10-19-16 11:23:22 am]  | CMD: cat /images/Win10/d1p1.img | /usr/local/sbin/udp-sender --interface eth0 --min-receivers 1 --max-wait 600 --mcast-data-address 192.168.10.60 --portbase 64158 --full-duplex --ttl 32 --nokbd --nopointopoint;cat /images/Win10/d1p2.img | /usr/local/sbin/udp-sender --interface eth0 --min-receivers 1 --max-wait 10 --mcast-data-address 192.168.10.60 --portbase 64158 --full-duplex --ttl 32 --nokbd --nopointopoint;
[10-19-16 11:23:22 am]  | Task (16) Multi-Cast Task has started.
[10-19-16 11:23:32 am]  | Task (16) Multi-Cast Task is already running PID 10132
[10-19-16 11:23:42 am]  | Task (16) Multi-Cast Task is already running PID 10132
[10-19-16 11:23:52 am]  | Task (16) Multi-Cast Task is already running PID 10132
[10-19-16 11:24:02 am]  | Task (16) Multi-Cast Task is already running PID 10132
[10-19-16 11:24:12 am]  | Task (16) Multi-Cast Task is already running PID 10132
[10-19-16 11:24:22 am]  | Task (16) Multi-Cast Task is already running PID 10132

Help will be appreciate, thanks
Matthieu

[Matthieu Jacquart]

Please help… ?
Everything seems fine when testing, just this screen remaining on Partclone screen

[Tom Elliott]

@Matthieu-Jacquart I believe this is related to the ports. I don’t know if you have a specific port or set of ports that are allowed vs, those that are not allowed.

[Wayne Workman]

@Matthieu-Jacquart said in Multicast remains on Partclone screen (HP switch):

Thanks for your help, deploying tens of computer in unicast is really long…

In the meantime, I would suggest only deploying 2 or 3 at a time. You can limit the number in Storage Management -> node -> max clients.

We need to figure out where the failure is exactly. Do you have a spare HP 2920/2530 switch that you can tinker with for a few days off-network ?

[Matthieu Jacquart]

@Wayne-Workman For the moment I just test multicast on one computer.
I have a 2530 spare to make test, but even if I try to deploy directly on my core L3 (2920), it’s similar…

[Wayne Workman]

@Matthieu-Jacquart My thought was to take two spare computers aside, install FOG on one with DHCP too, then take it off-network with your spare switch. And we just work on adjusting the switch configuration until we get multicast working to the other computer.