tRio skipping - frequent 'resend requests'

I'm running the latest JReceiver / tRio on 3 rebadged Dell receivers. JRec and Riot are running on a PIII-850 w/ 128MB RAM and Redhat 7.2.

Everything had been running great, but lately I'm noticing a lot of skipping and occasional dropouts when running Queue N Play mode. Looking at the the tRio log, I'm seeing a lot of re-send messages, e.g.

'Got resend request. tuneID: 2699, streamOffset: 10576031, reqOffset: 14912, reqLen: 9637'

I get ~10 resend requests per minute, though I only hear the skipping 2 - 3 times per song. Problem occurs on all 3 receivers.

Any ideas where the problem lies?

Are you using unicast, broadcast, or multicast as the send method in the trio.xml?

I have recently become suspicious of packets larger than the ethernet MTU not being sent as expected on broadcast sockets. I don't have a conclusive theory yet, though. I thought large UDP messages got fragmented into as many IP packets as necessary, but this may not always be the case. What I've been seeing is the receiver not getting a packet, and not sending a retry either.

-dB

For what it's worth, I had the same problem and eventually tracked it down to a faulty, brand-new, ebuyer 5port switch. After replacement the problem disappeared.

Good luck

The trio.xml file indicates I'm using multicast mode. While I'm fairly technically proficient, but Linux and networking are still daunting to me.

It seems the problem may be in the network hardware. I tried running with just the server, a 10/100 switch, and a couple of Rios- everything seems to work fine. But when I add the rest of the network (including DSL router, 2 wireless APs, and my working PC), things work fine at first but eventually start skipping again. I'm still playing with the network hardware components to try to isolate the problem to a switch or node.

Try the "unicast" send method in the trio.xml.

I've been talking with Paul (Mr Trio), and there are clearly issues in the
broadcast /multicast space, and in the use of larger UDP messages. Some are software related, some are hardware, and some are just that some of this stuff is enough out of the mainstream of network usage that it isn't very well stabilized

"If it worked, we wouldn't call it high-tech!"

Expect some active updates in the next month or so, I'd guess.

-dB