|
Title: GuiMiner - Use on Local Network Failed to Connect. And Too many Sockets? Post by: windygamer on May 24, 2011, 05:04:43 AM After a lot of searching and false turns, I finally found GuiMiner to make use of my two computers both with NVidia cards. So Thumbs up on a reasonable program for using GPUs for mining. ;D
I know I'm bucking the trend here, but I want to Solo mine on my two machines and point them both to the same BitCoin instance running on one of the machines to collect any coins in one account. I've got it running locally on one machine which has a GTX 470. And it connects to localhost on that machine to an instance of BitCoin.exe. That seems to work fine and I'm getting 84MH. Which we can talk about later if that seems like a problem. The other machine has a GTX 260, so I'd like to fire it up too. But I can't connect between the machines. Both machines are running Win 7 x64. I've opened the firewall for BitCoin and for guiminer.exe and for poclbm.exe for both Home/work(private) and Public networks. I'm guessing that I don't need to open it for guiminer.exe, but at this point I'm not sure. Using TcpView ( from here if you've not seen it: http://technet.microsoft.com/en-us/sysinternals/bb897437 ) I see that poclbm.exe is trying to connect to the machine running BitCoin. When I start the guiminer the following appears in the guiminer console: Quote 2011-05-23 21:47:33: Running command: poclbm.exe --user=uuuuu --pass=ppppp -o cyberstorm -p 8332 -d0 --verbose 2011-05-23 21:47:33: Listener for "Default" started 2011-05-23 21:47:43: Listener for "Default": 23/05/2011 21:47:43, Problems communicating with bitcoin RPC 2011-05-23 21:47:54: Listener for "Default": 23/05/2011 21:47:54, Problems communicating with bitcoin RPC 2011-05-23 21:48:05: Listener for "Default": 23/05/2011 21:48:05, Problems communicating with bitcoin RPC 2011-05-23 21:48:16: Listener for "Default": 23/05/2011 21:48:16, Problems communicating with bitcoin RPC Oddly enough. Using TCPView on the local machine shows this: bitcoin.exe:2464 TCP AlienStorm:8333 AlienStorm:0 LISTENING How strange. The local machine appears to be listening on 8333. Is this because 8332 is used on the other machine, and they both need to be unique over the public network since I'm behind a NAT router? So to connect to my local machine I have to use port 8333. Ok I guess I can deal with that. But I'd like to use the other machine on the network. Any clue why guiminer [poclbm.exe] can't connect to cyberstorm? Actually, it does not appear to be able to connect to the local machine either. TCPview shows a pile of strange entries: Quote [System Process]:0 TCP AlienStorm:8332 localhost:55859 TIME_WAIT [System Process]:0 TCP alienstorm:56297 192.168.1.1:65535 TIME_WAIT [System Process]:0 TCP AlienStorm:56296 localhost:8333 TIME_WAIT [System Process]:0 TCP alienstorm:56300 192.168.1.1:60001 TIME_WAIT [System Process]:0 TCP AlienStorm:56299 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56302 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56306 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56309 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56311 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56313 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56316 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56319 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56321 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56323 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56328 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56332 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56334 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56336 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56340 localhost:8333 TIME_WAIT [System Process]:0 TCP AlienStorm:56342 localhost:8333 TIME_WAIT bitcoin.exe:2464 TCP AlienStorm:8333 AlienStorm:0 LISTENING Not knowing any better, I would guess that someone is not cleaning up their failed sockets. But I do see this: Quote poclbm.exe:1440 TCP AlienStorm:56344 localhost:8333 ESTABLISHED Which leads me to guess that the problem is not that the connection can't be established, but that the protocol fails in some other way. Thanks for your help. WG Title: Not clear what changed... But local works Post by: windygamer on May 24, 2011, 06:45:15 AM Not clear what changed, but now localhost works with the same local server.
Does the Bitcoin server need to complete loading the blocks or something before mining succeeds? Seems like there was a communication error, and now it's working fine. Locally. 39 Mh/s, which is about what I'd expect from a GTX 260. I still can't connect to the remote host to mine into a single instance of bitcoin.exe on my network. Thanks, wg Title: Re: GuiMiner - Use on Local Network Failed to Connect. And Too many Sockets? Post by: foggyb on May 24, 2011, 05:32:37 PM Same problem here: Server is Win7 x64 SP1 and client is Win7 x32 with no updates.
I'm going to try XP next on client. I noticed some weird behaviour also. Locally I was able to get GUIminer working, but then I tried to repeat my results, I can no longer connect to local bitcoin server.' GUIminer client on remote client PC is stuck in "connecting" mode. Here is my TCP/IP scan: http://dl.dropbox.com/u/5765587/net.png Seems to show local GUIminer is connected at 8332, but GUIminer doesn't think so. Title: And a related question Post by: windygamer on May 25, 2011, 05:03:47 AM After running for a couple of days I see:
Accepted 1775 (68) On the summary page. Any clue what this means? Something like 1776 hashes found, and 68 are still being worked on? Stale is just a -, which I'm guessing means that these have been claimed by someone else, or are otherwise unavailable. Also, any guidance about how long before I might actually see a coin or two? It would seem from these results that I'm repeating a large number of solutions that other folks have already claimed, so I'm not getting any coins. I'm just trying to confirm that things are actually working. The main bitcoin display is completely devoid of any useful display at this point. Since I'm not able to connect the two machines to the same bitcoin instance, but rather there are two instances, I presume that the two systems are solving the same hashes, which is certainly sub-optimal. Am I correct in assuming that if the two machines were talking to the same instance then they would divide the work without duplication? Thanks, wg |
