Comment 3 for bug 1246128

Raghu, was it a debug build? Sadly

> # 2013-10-29T10:10:35 [18883] ==00:00:00:24.034 20352== at 0x650919A: gcomm::GMCast::handle_up(void const*, gcomm::Datagram const&, gcomm::ProtoUpMeta const&) (gmcast_proto.hpp:119)

must refer to either of these 2 lines in *gmcast.cpp*:handle_up(), but it would be great to know for sure:

           if (msg.type() >= Message::T_USER_BASE)
            {
                if (msg.flags() & Message::F_RELAY)
                {
                    relay(msg,
                          Datagram(dg, dg.offset() + msg.serial_size()),
                          id);
                }
                send_up(Datagram(dg, dg.offset() + msg.serial_size()),
                        ProtoUpMeta(msg.source_uuid()));
                p->set_tstamp(gu::datetime::Date::now());
            }
            else
            {
                try
                {
                    gu_trace(p->handle_message(msg));
                    p->set_tstamp(gu::datetime::Date::now());
                }

was the node shut down under load? Is it the only occurrence reported by valgrind?