Overloaded my router with mining?

[Yanz]

So I just added another 2 gpus to my mining farm to a total of 7 gpus now. Funny thing is when I added that system, everything went haywire. cgminer (1.6.2) started lagging out and saying that Pool X is not providing work fast enough. I thought it was my pool so I added some more. They all lag out at the same time and then shortly after they all recover. I've tried different routers and my Internet seems fine. Ping times are about 100ms to btcguild and 150ish to bitcoins.lc. I've tried eligus as well but it seem to lag out too. The new system I added was Ubuntu natty and the others are running Linuxcoin final. Is this an OS issue or what?

[1714988675]

1714988675

[1714988675]

1714988675

[1714988675]

1714988675

[TeaL]

I'm almost certain it's not your available bandwidth. I monitored my network connections while mining and it doesn't appear to use much data at all. I'd be guessing it's a software problem on your end, what that is, I don't know.

[MelMan2002]

Could also be a heat issue - but really hard to tell with limited information.

[deslok]

I have 3 miners on a wireless network i've seen an error message after a restart about an ip address conflict, try assigning either static or reserved ip's and see if that helps

[Yanz]

Yep. It's not a router problem.
I don't think it's a heat issue. GPUs never go above 75C. I have plenty of power.
Don't think it's an ip as they all get ip's from dhcp but I'll double check. Funny is that it works for a while, like a minute then lags out. And then my script will check that the gpus arent loaded and restart the miners. Everything repeats.

[3phase]

Well, back to basics:

If you can do a ping test (like 1000 packets) from the router to your destination pool successfully (not more than 1 packet lost) then your Internet connection is OK and you can try the following:

1. Change the Ethernet cable from the miner to the switch and the cable from switch to router (one-by-one)
2. Make sure it is not getting intense interference from any parallel running power cord very close to it
3. Try a different switch in case the one you have has problems
4. Connect only the miners to the switch in case another PC on the net is broadcasting packets and causing congestion, and then connect only that switch to the router.
5. Connect each miner one-by-one to the switch and connect the switch to the router.
6. Connect each miner one-by-one to the router directly

In each of these cases try a "ping -n 1000" (windows) or "ping -c 1000" (linux) from the miner to the router and you should get 1000 replies.

Work out what causes the problem from the results.

Plain old Ethernet debugging

Good luck

[Yanz]

Not sure what caused this but it went away on its own...weird.
Thanks, 3phase I tried most of that already. Except for the power cord...all my cords run together in a nice little bundle...

[c_k]

If you get it happening again, try the -Q option in cgminer - start at 1 and steadily make your way up to a max of 10 or if no luck there try -G (could be a lowercase g, I can't remember off the top of my head) and 1 so you're only using 1 thread per GPU.

Also, the latest version is 2.0.2, you definitely sent to upgrade IMO