Problem With Proxy & Bfgminer?

[OliRS]

Setup proxy went fine on my Model B 512MB Raspberry Pi, runs fine and can mine to Slush's pool via the proxy running on it, but I regularly get this message in the proxy output:

2014-02-11 20:16:06,885 INFO proxy client_service.handle_event # New job 138db for prevhash a246d9bd, clean_jobs=False 2014-02-11 20:16:07,004 WARNING proxy stratum_listener.get_transactions # mining.get_transactions isn't supported by proxy

As soon as this appears in the proxy output, in Bfgminer I get this:

[2014-02-11 20:16:07] Stratum from pool 0 requested work update [2014-02-11 20:16:07] Pool 0 is sending mismatched block contents to us (0 is not 512-1023)

The number range at the end varies.

If I mine with Bfgminer directly to the pool rather than via the proxy it doesn't show any errors regarding mismatched block contents.

Is this something wrong or an error not being handled correctly or something? Bit concerned I have done something wrong.

Thanks.

[flound1129]

Also getting this error with bfgminer, latest version.

[daryltucker]

This is happening to me, as well. It seems to become progressively worse, to the point that I am no longer mining anything after an hour or so.

https://github.com/slush0/stratum-mining-proxy/issues/77

[ab1jx]

Because somebody will stumble across this thread eventually like I did, this is what was causing it for me. The clue was that I restarted the proxy and at the beginning of its output I saw something about a bad HTTP request. I was flailing around last night and in my bfgminer.conf I put in http://pi2:3333 as a URL rather than using the stratum+tcp://pi2:3333 specialized stratum protocol. If you have several servers in your pool bfgminer will probably only choose this one part of the time. And the error messages kept coming even after I gave up for the night and went back to cpuminer. The next morning I rebooted the worker machine and that stopped them. cpuminer was working all night, I just had these messages in the proxy output. They apparently get queued up, I don't know if they would have stopped by themselves or not.

OK, that didn't fix it after all, but it was caused by bfgminer. I'll reboot and try cgminer. Half a day later I've been running a fork of cgminer by dmaxl from https://github.com/dmaxl/cgminer/ with great success. It's been running for several hours, seems stable, I can adjust my ASIC's clock up and down, I haven't gotten any API to work yet. But there's a ton of stuff out there, I've tried several versions each of bfgminer, cgminer, and even hacks on cpuminer to let it drive Gridseed ASICs. Nothing really wasted but time, which could be chalked up to education.

And then I stopped using a proxy altogether because you can only have 1 difficulty level per proxy and I was doing a mix of cpu mining and using an ASIC. I gave up on bfgminer though, just used cgminer.

[eliddell1]

my pi won't seem to make that version of cgminer i get the error:

gcc: error: ’-I/usr/include/curl’: No such file or directory
Makefile:859: recipe for target 'cgminer-cgminer.o' failed
make[2]: *** [cgminer-cgminer.o] Error 1
make[2]: Leaving directory '/root/cgminer'
Makefile:1427: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/root/cgminer'
Makefile:682: recipe for target 'all' failed
make: *** [all] Error 2

[ab1jx]

Look at the first line, you don't have curl probably. Or it's in a different place.

If you have locate, find curl.h and set your -I path to where it is instead of /usr/include/

Oh, pi, yes. On my pi it's at /usr/include/curl/curl.h so you don't have curl. There's probably other stuff missing too. Do sudo apt-get install -y build-essential git libusb-1.0-0-dev libusb-1.0-0 libcurl4-openssl-dev libncurses5-dev libudev-dev screen libtool automake pkg-config libjansson-dev

That's from https://bitcointalk.org/index.php?topic=1764803.0 for a different cgminer version but it should work. Anything you've already got installed it'll just keep.