https://community.hpe.com/t5/bladesystem-general/multicast-and-blade-systems/m-p/4305557#M4388 Hi,<BR /><BR />We have experienced some strange issues when we run some multicast tests with two setups:<BR /><BR />1. C7000 enclosure, bl685c Blades, GB2ec intraconnect switches. Red Hat 5.2, one of the servers has mrouted installed as a igmp routing daemon. (I have understood that the GB2ec switches can't act as one)<BR /><BR />2. Same as above but with Virtual Connect modules instead of GB2ec, connected to Extreme switches which acts as IGMP routers.<BR /><BR /><BR />Generally we have noticed that the servers that sends UDP traffic has very high load on the CPU, but the receivers seem to be okay.<BR /><BR />The strangest issue, which is our real problem, happened when we tried some tests on the configuration 1 above.<BR />When doing multicast tests, we can't get higher speed than 100Mbit/s as a total within the enclosure.<BR /><BR />We first tried to send multicast from server A to server B.<BR />The result was that server A had a transfer rate of 100MBIT and the listener as well.<BR /><BR />We then tried to send multicast from server C to server D at the same time.<BR />The result was that both server A and server C had a transfer rate of 50MBIT.<BR /><BR />It seems like there is some limitation within the enclosure/switches that limits this for us? Thu, 13 Nov 2008 11:21:48 GMT Linus Hedström 2008-11-13T11:21:48Z https://community.hpe.com/t5/bladesystem-general/multicast-and-blade-systems/m-p/4305557#M4388 Hi,<BR /><BR />We have experienced some strange issues when we run some multicast tests with two setups:<BR /><BR />1. C7000 enclosure, bl685c Blades, GB2ec intraconnect switches. Red Hat 5.2, one of the servers has mrouted installed as a igmp routing daemon. (I have understood that the GB2ec switches can't act as one)<BR /><BR />2. Same as above but with Virtual Connect modules instead of GB2ec, connected to Extreme switches which acts as IGMP routers.<BR /><BR /><BR />Generally we have noticed that the servers that sends UDP traffic has very high load on the CPU, but the receivers seem to be okay.<BR /><BR />The strangest issue, which is our real problem, happened when we tried some tests on the configuration 1 above.<BR />When doing multicast tests, we can't get higher speed than 100Mbit/s as a total within the enclosure.<BR /><BR />We first tried to send multicast from server A to server B.<BR />The result was that server A had a transfer rate of 100MBIT and the listener as well.<BR /><BR />We then tried to send multicast from server C to server D at the same time.<BR />The result was that both server A and server C had a transfer rate of 50MBIT.<BR /><BR />It seems like there is some limitation within the enclosure/switches that limits this for us? Thu, 13 Nov 2008 11:21:48 GMT https://community.hpe.com/t5/bladesystem-general/multicast-and-blade-systems/m-p/4305557#M4388 Linus Hedström 2008-11-13T11:21:48Z