> But the problem is when one of the node gets power-cycled, the receive > qp never gets notified about the other end dying and it just hangs > around forever. Is there a keep-alive that is maintained between the > connection managers of the two nodes ? Is there a timeout that can be > set on the qp so that these kinds of events get notified fast. ?
No, if you don't explicitly send anything then nothing will be sent automatically. So if you want keep-alives, you will need to implement them at the application level. - R. _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
