[evla-sw-discuss] Baseline Boards to mccctest

[Barry Clark]

Any chance there is something else running on mccctest that is
gobbling up the packets?  I believe if the packets are addressed
to a specific socket, and there are two copies receiving, they
only go to one.

On 04/30/2014 04:23 PM, Kevin Ryan wrote:
> As Bill mentioned at the Wed. Meeting this morning, lag frame data from the Baseline Boards are no longer making it into CRM running on mccctest.
>
> We did some testing and here is what we found:
>
> Test 1 -- Baseline Board packets to mccc and mccctest:
>
> Bruce and I set up a baseline board to transmit lag frames to various ports and IPs.  A little UDP receiver program was made in Java to simulate the CRM and was run on both 'mccc' and 'mccctest'.
>
> The 'netcat' Linux command (thanks Bruce) was also used to listen for UDP packets from the Baseline Board so as to eliminate a possible Java/JVM issue.
>
> The results are that we are able to see packets from the Baseline Board at 'mccc' using both Java and 'netcat' and we are _not_ able to see Baseline Board packets at 'mccctest' using either Java or 'netcat'.
>
> Test 2 -- UDP from a CBE node to mccc and mccctest:
> A little 'Hello World' UDP transmitter was run on cbe-node-01 (so it would be on the correct network).  It was received by both mccc and mccctest using both netcat and Java programs.
>
> Since things were a bit 'ify' on the system today (boards being pulled, CRM running and whatnot) and we had to stop testing so Ken could recover the system before too late, I would like to try this all again tomorrow just to be sure of the findings.
>
> But as it stands now - mccctest can receive, on the secondary network interface, using Java, a UDP packet derived in Java from a CBE node but cannot read a UDP packet from a Baseline Board on the same interface.
>
> Kevin
>
>
> _______________________________________________
> evla-sw-discuss mailing list
> evla-sw-discuss at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evla-sw-discuss