[e-nsp] multicast packets reordering

Marcin Kuczera marcin at leon.pl
Tue Sep 23 13:54:00 EDT 2008


Fabian wrote:
> Hi Marcin,
> 
> It should not (and if using LAG/Multilink with 'i' Extreme switches, all
> Mcast traffic goes through the same port).
> 
> Are you using them with IP Mcast routing (PIM, DVMRP) or simple Layer 2
> VLAN?
> I would suggest to test without the Redback or connecting your TestSystem
> level by level.

ok, little correction.
Signal provider has removed some flapping MAC (router) from his network, 
so now out-of-order packets look like that:
Sep 22 15:29:15 10.200.180.4 iperf: [ 3] 16600.0-16610.0 sec 1 datagrams
received out-of-order
Sep 22 15:38:05 10.200.180.4 iperf: [ 3] 17130.0-17140.0 sec 1 datagrams
received out-of-order
Sep 22 15:47:05 10.200.180.4 iperf: [ 3] 17670.0-17680.0 sec 1 datagrams
received out-of-order
Sep 22 15:55:55 10.200.180.4 iperf: [ 3] 18200.0-18210.0 sec 1 datagrams
received out-of-order
Sep 22 16:04:45 10.200.180.4 iperf: [ 3] 18730.0-18740.0 sec 1 datagrams
received out-of-order
Sep 22 16:13:45 10.200.180.4 iperf: [ 3] 19270.0-19280.0 sec 1 datagrams
received out-of-order
Sep 22 16:22:35 10.200.180.4 iperf: [ 3] 19800.0-19810.0 sec 1 datagrams
received out-of-order
Sep 22 16:31:25 10.200.180.4 iperf: [ 3] 20330.0-20340.0 sec 1 datagrams
received out-of-order
Sep 22 16:40:15 10.200.180.4 iperf: [ 3] 20860.0-20870.0 sec 1 datagrams
received out-of-order


exaclty every 530seconds !
connectivity is like that:
source-->AlpineA-->RedBack(PIM)-->AlpineA-->AlpineB-->Summit48-->TestSystem.

soft in Alpine 7.6.4.4
soft in RedBack 6.1.1.4


Any Idea what could that be ?

Regards,
Marcin







> 
> I'am testing IP TV using extreme core based on quite old, but
> still alive Alpine3808,3804 and Summit5i switches.
> 
> At the moment, the test device generating multicast reports
> is connected in this way:
> 
> source-->AlpineA-->RedBack(PIM)-->AlpineA-->AlpineB-->Summit48-->TestSystem.
> 
> TestSystem reports:
> Sep 18 17:41:00 10.200.180.4 iperf: [ 3] 11310.0-11320.0 sec 2 datagrams
> received out-of-order
> Sep 18 17:41:30 10.200.180.4 iperf: [ 3] 11340.0-11350.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:42:10 10.200.180.4 iperf: [ 3] 11380.0-11390.0 sec 2 datagrams
> received out-of-order
> Sep 18 17:42:40 10.200.180.4 iperf: [ 3] 11410.0-11420.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:43:00 10.200.180.4 iperf: [ 3] 11430.0-11440.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:44:20 10.200.180.4 iperf: [ 3] 11510.0-11520.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:45:40 10.200.180.4 iperf: [ 3] 11590.0-11600.0 sec 2 datagrams
> received out-of-order
> Sep 18 17:45:50 10.200.180.4 iperf: [ 3] 11600.0-11610.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:46:00 10.200.180.4 iperf: [ 3] 11610.0-11620.0 sec 3 datagrams
> received out-of-order
> Sep 18 17:46:10 10.200.180.4 iperf: [ 3] 11620.0-11630.0 sec 1 datagrams
> received out-of-order
> Sep 18 17:46:30 10.200.180.4 iperf: [ 3] 11640.0-11650.0 sec 1 datagrams
> received out-of-order
> 
> looks nasty...
> 
> soft in Alpine 7.6.4.4
> soft in RedBack 6.1.1.4
> 
> No multilinks...
> 
> Is there any possibility that Alpine does the reordering of packets ??
> If yes, is there any solution to that problem ?
> 
> Reordering is common problem from the source side, but... I have 10x 
> more reorders that others..
> 
> Regards,
> Marcin
> _______________________________________________
> extreme-nsp mailing list
> extreme-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/extreme-nsp
> 
> 



More information about the extreme-nsp mailing list