Interface eth0 dies when transfering video.

[omnius]

I've got a hacked SA AT&T S2 running 3.0 with a DLink USB network adaptor hooked up to it. Everything appears fine, it grabs an IP and gets its guide data over the network. However, if I do anything extensive on the network connection like extract video with something like tserver or mfs_ftp, it will go fine for a while at around 600kB/sec, but will then just suddenly stop. After it dies, it no longer responds on the network at all. I can't ping the tivo and the tivo can't ping my PC. Nothing else like telnet, ftp, www, can get to the tivo. Typing "ifconfig" on the tivo reveals that the eth0 interface has dissapeared completely (it only shows the loopback lo). I have to reboot the Tivo to get eth0 back up and running.

I usually get around 600-800 MB before it dies, which is not nearly enough to get even a 1 hour TV show. So video extraction basically is useless for me until I figure this out.

Anybody have any ideas or have experienced a problem like this? I can't try another USB network card without buying another one. Also, both of the programs I've tried were using the same S2 compiled mfs_stream that came with tserver_mfs_mips6 posted by jdiner.

[omnius]

Well, I take back part of that. I can use ifconfig to bring it right back up "ifconfig eth0 ipaddress up" and it works again. But there's no reason for it to go down like it is. Sometimes it only transfers around 50mb before it dies. I've also tried limiting the speed to half (300kB/s) and it still dies.

[fixn278]

I don't know if this will help you any, but I have found Dlink products to be extremely unreliable. If you can get your hands on another brand of adapter to test, I would recommend it.

Keep in mind, I haven't touched S2 USB Ethernet, this is just a person product-quality observation.

[omnius]

The DLink card itself was one of the first things I thought of, but I've actually used DLink extensively and never really had any problems. I'll probably grab a laptop or something and hook the dlink up to it and put it through its paces to find out if thats the problem.

[omnius]

I figured out a bandaid fix. It turns out that as long as I bring eth0 back up before the transfer's timeout happens, it picks it right back up and keeps on transfering. So I just wrote a little shell script that checks every 10 seconds to see if eth0 is up or not, and brings it back up if its down.

This is a question for you gurus out there. Is there a possible reason besides a troubled network connection for ifconfig to be reporting this many errors on eth0?

inet addr:192.168.0.3 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:29184 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:87535478 dropped:58115 overruns:1 carrier:0 coll:0

Thats showing 0 packets transmitted but a whole crapload of errors and drops. Is that perhaps a bug and its switching the transmitted packects with the errors? Since I'm sending large files from the tivo to my computer, there should be a heck of a lot more than 0 transmitted packets.

[BubbleLamp]

Originally posted by omnius
I figured out a bandaid fix. It turns out that as long as I bring eth0 back up before the transfer's timeout happens, it picks it right back up and keeps on transfering. So I just wrote a little shell script that checks every 10 seconds to see if eth0 is up or not, and brings it back up if its down.

This is a question for you gurus out there. Is there a possible reason besides a troubled network connection for ifconfig to be reporting this many errors on eth0?

inet addr:192.168.0.3 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:29184 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:87535478 dropped:58115 overruns:1 carrier:0 coll:0

Thats showing 0 packets transmitted but a whole crapload of errors and drops. Is that perhaps a bug and its switching the transmitted packects with the errors? Since I'm sending large files from the tivo to my computer, there should be a heck of a lot more than 0 transmitted packets.

You have a serious problem there. You should never see that many errors. You ether have a bad or crappy ethernet cable, bad NIC/HUB/SWITCH, or you are trying to run full duplex on a half-duplex network.

[pcompact]

This is off-topic. But I needed to find a hackable SA series2 backup.
Thank you