Sure but the multicast packets on the primary (public) interface (ce0) seem to be duplicated by the solaris ip stack while they are not duplicated on the second dedicated interface of the sunray interconnect (ce1). I would expect the SRS multicast packets to be heartbeats and single messages at regular intervals and not being duplicated on the same interface each time one is sent on the primary interface. The main issue is that this behavior seems to be causing our switch to drop unicast traffic regularly after multicast messages are sent on the primary interface. This may be a bug in the switch but why should the SRS software send two consecutive multicast messages containing identical data and with nearly identical timestamps while it only sends single multicast messages on the second interface as expected ?
Snoop traffic on ce0: ________________________________ 38 0.56873 sr1-ce0 -> 224.101.101.101 ETHER Type=0800 (IP), size = 367 bytes 38 0.56873 sr1-ce0 -> 224.101.101.101 IP D=224.101.101.101 S=xxx LEN=353, ID=5658, TOS=0x0, TTL=1 38 0.56873 sr1-ce0 -> 224.101.101.101 UDP D=7009 S=7009 LEN=333 ________________________________ 39 0.00033 sr1-ce0 -> 224.101.101.101 ETHER Type=0800 (IP), size = 367 bytes 39 0.00033 sr1-ce0 -> 224.101.101.101 IP D=224.101.101.101 S=xxx LEN=353, ID=5658, TOS=0x0, TTL=1 39 0.00033 sr1-ce0 -> 224.101.101.101 UDP D=7009 S=7009 LEN=333 Snoop traffic on ce1: 7 0.42615 sr1-ce1 -> 224.101.101.101 ETHER Type=0800 (IP), size = 367 bytes 7 0.42615 sr1-ce1 -> 224.101.101.101 IP D=224.101.101.101 S=172.16.128.7 LEN=353, ID=59077, TOS=0x0, TTL=1 7 0.42615 sr1-ce1 -> 224.101.101.101 UDP D=7009 S=7009 LEN=333 Could you try to snoop for multicast messages on your primary interface to see if you are getting the same behavior as me ? Thanks Thierry. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Kent Peacock Sent: 05 September 2008 18:32 To: SunRay-Users mailing list Subject: Re: [SunRay-Users] multicasting problems with sunray software andsolaris On 09/05/08 10:25, Thierry Delaitre wrote: > Hello, > > > > I have some multicasting issues with the HP switch we are using. When > one of the 2 sunray server in the failover load-balanced group is > sending a multicast message, then the switch is dropping unicast packets > to that host for 10/20 seconds L Disabling IGMP solves the problem but > it is not really ideal. > > > > When we are using wireshark or snoop on the solaris 10 05/08 (with SRS4) > sunray server, we can observe that Solaris is duplicating the sunray > multicast messages on the primary public interface!! Is this normal ? > This strange behavior could be triggering a bug on our HP switch. The Sun Ray server software sends multicast heartbeat packets on all of the real interfaces on the host. It's how it discovers the connectivity of the servers. Kent _______________________________________________ SunRay-Users mailing list [email protected] http://www.filibeto.org/mailman/listinfo/sunray-users _______________________________________________ SunRay-Users mailing list [email protected] http://www.filibeto.org/mailman/listinfo/sunray-users
