Topic: CGWatcher 1.4.0, the GUI/monitor for CGMiner and BFGMiner to prevent downtime - page 14. (Read 180386 times)

Taxidermista:  Please email me again because my inbox has been flooded over the past few weeks. There are around 200 people in the beta and of the few that reported issues connecting, it has always been a firewall issue or a port-forwarding issue. You look to have port-forwarding done correctly, so my guess would be a firewall issue. When you email me, include the bug report created by clicking 'Report Bug' in the toolbar.


Einewton:  Yes, a web application and Android app are planned, but not until the desktop application is released. The web application will be located at http://MinerRemote.com. I've received some offers for help on these, so development should be pretty quick.

I know that CG Remote is a standalone app, but it would also be cool if you could expose it, and have a web version so that I can control my miners from work/cell/anywhere else... I guess my question would be, are there plans to provide a web interface with CG Remote?

If not.... Could you provide an API (JSON, Soap, Whatever) or some other method that I can make my own CG Remote WebApp?

Sorry, I posted here because I asked you by mail but didn't get any answer.

I've tried enabling that option in the Windows Firewall but no success, still not responding.

Taxidermista:  Please email me when it is related to CGRemote until I create a forum post for it (which I'll do closer to release). Also, include the bug report (click Report Bug in toolbar) when reporting problems. If you have ports forwarded correctly, it is 99% likely a firewall issue. Make sure CGWatcher and CGRemote have access through their respective firewalls. If you're using Windows Firewall, you may need to create an additional 'Allow' rule for the port on each miner. I ran into this problem, but the computer was also running Windows Firewall Control, so that may have been the cause. For incoming firewall rules on the miners (for CGWatcher), enable "Allow edge traversal" in the Advanced tab of the rule's properties (using Windows Firewall with Advanced Security). If you're using different firewall software, let me know what it is and I can try to help.


I'm still hoping to have the CGWatcher update out this weekend, although my time estimates are almost always wrong.

I don't know what's wrong but this conf does not work. I can't connect with my six miners & respectives cgwatchers forwarding every 5121 port to ports 6661 to 6666, I configure CGRemote with my public ip and these 6 ports, of course. CGRemote reports that the miners can't be reached. This is my router NAT table:



I've tried selecting protocol TCP/UDP in one of the miners but did not connect either.

Thanks for your detailed answer and of course this fantastic app of yours. I’ll wait for the update than. 

Yeah, I'd rather get a BSOD then give up the ADL beni's... You might as well OC your cards in MSI Afterburner instead of setting your clock speeds in the CGMiner application, that would also prevent BSOD.


--no-adl            Disable the ATI display library used for monitoring and setting GPU parameters


I'll be waiting for this weekends release to have it close the cmd, instead of the Quit method.

Thank you very much for taking the time to answer our questions. I'm going to forward those ports in my router and try a dynamic ip hostname service.

Related to the bsod when closing with the Quit command I've read in the Litecointalk forum that you can avoid the crash using the flag --no-adl in cgminer but I don't know if this is a practical solution or even worse than the problem you're trying to solve.

Your an animal.. Keep up the good work, man. I like where this product is going.

Taxidermista:

I'm going to make changes that will kill the miner rather than shut it down during restarts, which should help with stability.. at least when restarting using the same profile or a profile using the same algorithm.

As far as how CGRemote works, the short answer: magic. The long answer: port forwarding. If there wasn't a way to reach your computer from the Internet you would not have been able to post that reply. Port forwarding tells your router that when data comes in on a specific port, forward it to a specific computer on the LAN. Each computer on your LAN would then use a unique port, each setup for the router to forward. If you want to use a dynamic IP hostname service like dyndns you can, or you can just update the IP address(es) when/if they change.



Goldshredder:

https://bitcointalksearch.org/topic/m.3853830



frufru:

The reason I chose not to use the startup folder is because it seems to not work 100% of the time, so I'm not completely surprised that didn't work. I just got some Windows 8 PCs set up so I'll have to see what new unexpected behaviors are found with modifying the registry. CGWatcher should try to set itself to the HKLM/Software/Microsoft/Windows/CurrentVersion/Run key (all users), and if that fails it tries the HKCU/Software/Microsoft/Windows/CurrentVersion/Run key (current user only). However, if you're not running CGWatcher as Administrator, Windows virtualizes the HKLM hive so CGWatcher thinks it saved correctly but it was actually stored to some WOW6432Node key somewhere. I think what I need to do is just set CGWatcher to start with Windows on a per-user basis, so HKCU only. If you need it to start automatically for multiple users, just open it at least once in each user with the "start when Windows starts" option enabled.

This will be changed in the next update, which I'll try to have out this weekend.

Same problem here. I have two partions on my rig, one with win7 and one with a fresh win8 install. I have no problems on win7 but on the new system cgwatcher hangs when I try to save settings and won't start with windows. I moved the my mining folder from win7 so all the settings and files are the same. I also tried a fresh start but no change. For some strange reason cgwatcher won't start automatically even if I put the shortcut in the startup folder.

Darn I hope this isn't a silly question (the presumed common-ness of my setup makes me think it probably is...), but how does one run multiple instances of CGWatcher on one (Windows) machine, with each one watching a different running instance of CGminer?

i.e. I have a cgminer running for gpus and a different cgminer running for some asics, both at the same time.

When I run the second instance of CGWatcher so I can set it up to watch my asic version of cgminer, it just picks up the existing gpu miner that's running and displays all the info for that.  And setting up / changing profile in that second CGWatcher instance, seems to stop the gpu miner profile that's running in the first CGWatcher instance.  So no idea at this stage how to run both independently, help appreciated!

This Quit command crash is driving me crazy. It ruins half the good things you expect from CGWatcher and CGRemote. Changing mined coin on the fly from CGRemote was top priority for me and I can't do it. 

milone, could you please answer my email about using CGRemote over the internet? I don't understand how the program can know my public ip address to connect with miners on my local network. Do I have to use some service like myip.us or something?

Bingo! Setting gpu-threads to 1 in the config fixed it! Thanks!

Einewton:  I'm not sure why that would happen, but it's been a while since I've visited that part of the code so it's possible that I made changes recently that had unexpected affects on the monitor. I'll be looking into this the first chance I get. (Out of town for the next week or two and have to setup Visual Studio on my new notebook). The monitor has always been my first priority with CGWatcher, so if something is not working correctly on it I will make sure it gets fixed.


knowone:  If the miner has a lower hashrate after being started with CGWatcher, it is most likely due to the gpu-threads setting. In cgminer's readme it states that the default value for gpu-threads is 2. So if you do not explicitly set the gpu-threads setting in arguments or config file, it should use 2. But it doesn't, it uses 1. This is something a user pointed out a while back and I tested myself and confirmed. So I made CGWatcher set it to 2 by default, so it would produce the expected behavior one would expect after reading cgminer's readme. This means that starting cgminer outside of CGWatcher uses 1 gpu thread, and starting cgminer with CGWatcher uses 2 threads. Whenever someone has brought up seeing a lower hashrate when the miner is started by CGWatcher, it has been because of this 100% of the time.

In CGWatcher you don't have to point to the .exe when configuring the profile. If you already have a batch file you can point to that instead. CGWatcher will load any arguments from the batch file, and if you change them in Profile Manager it will save them back to the batch file. It will not affect the setx commands or any other commands in the batch file, those will remain in place. I've seen conflicting statements on how long the setx commands are applied for - some say you only need to run the commands once, others that the command will work until the next computer restart. I don't know which is true, and I never really took the time to find out. But if you would prefer running them before launching the miner - just in case - then I would recommend pointing the profile to the batch file instead of the executable.

I want to use CGWatcher, but I find that I get a much lower hashrate when I start cgminer from CGWatcher than from a batch file.

Example batch file:

ECHO This is "GPU mining.bat".
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1

cgminer --config mdl.conf

pause
exit
===============
Example config file = mdl.conf:

{
"pools" : [
   {   "url" : "stratm+tcp://middlecoin.com:3333",
      "user" : "redacted",
      "pass" : "redacted"
   },
   {   "url" : "stratum+tcp://ltc.give-me-coins.com:3334",
      "user" : "redacted",
      "pass" : "redacted"
   }
   ],
"usb" : ":0",
"scrypt" : true,
"gpu-platform" : "1",
"thread-concurrency" : "21712",
"auto-fan" : true,
"auto-gpu" : true,
"gpu-engine" : "900-1150",
"temp-target" : "80",
"temp-overheat" : "90",
"temp-cutoff" : "95","intensity" : "13",
"worksize" : "256",
"api-listen" : true,
"api-allow" : "W: 127.0.0.1"
}
============

(The same problem also happens with a different config using a single pool.)

I am using cgminer v. 3.6.6, CGWatcher 1.3.3.2, Windows 7 Pro 64 bit SP 1. When cgminer is already running and I start up CGWatcher, it monitors okay and does not affect the hash rate. When starting from the batch file, both cgminer and CGWatcher report a hash rate around 560 Kh/s. When starting or restarting from CGWatcher, I usually get 5 to 20 Kh/s, although once I got in the 450 range.

I hate to report the same problem someone else did, but in a comment at http://manotechnology.blogspot.com/p/cgwatcher.html, Tomáš Bräuer on August 14, 2013, reported his rate dropping from around 800 Kh/s to 620 Kh/s under the same circumstances. Unfortunately, for some reason I can't load the next page of comments from that website to see what he was told.

One oddity is that when monitoring, CGWatcher reports:
[12/9/2013 4:35:05 PM]       -- Begin CGWatcher v1.3.3.2 Process -------------------------------------------------------
[12/9/2013 4:35:09 PM]       Running miner detected: CGMiner using "NEW Fastcoin" profile.
[12/9/2013 4:35:09 PM]       Active profile is "NEW Fastcoin"
[12/9/2013 4:35:09 PM]       CGMiner instance found.
[12/9/2013 4:35:09 PM]       The current running miner's profile is "NEW Fastcoin".
[12/9/2013 4:35:10 PM]       Monitoring is turned on at 30 second intervals.
====================
       . . . and "NEW Fastcoin" profile points to a different pool than the one cgminer is using.

After collecting the log above, I paused and then resumed mining and it came up with the Fastcoin config, running at a 10 Kh/s rate. I then checked the "Show debug log entries" box, paused and restarted cgminer, and the new rate is about 430 Kh/s (vs over 500 when starting from the batch file).

The whole log now reads as follows:
[12/9/2013 4:35:05 PM]       -- Begin CGWatcher v1.3.3.2 Process -------------------------------------------------------
[12/9/2013 4:35:05 PM]   [d] OpenCL.Initialize Error 5: Unable to get device info for device 490231200
[12/9/2013 4:35:06 PM]   [d] WER DontShowUI value exists (1) and was set to 1 (disabled).
[12/9/2013 4:35:08 PM]   [d] 2 miner instances found, 2 ignored.
[12/9/2013 4:35:09 PM]       Running miner detected: CGMiner using "NEW Fastcoin" profile.
[12/9/2013 4:35:09 PM]   [d] Miner.IsProcessAlive(1476) returns True. [First try, process is alive.] (CPID1476/CMPID=0/CBPID=0)
[12/9/2013 4:35:09 PM]       Active profile is "NEW Fastcoin"
[12/9/2013 4:35:09 PM]       CGMiner instance found.
[12/9/2013 4:35:09 PM]       The current running miner's profile is "NEW Fastcoin".
[12/9/2013 4:35:09 PM]   [d] Coins.Refresh complete, 54 coin(s) present.
[12/9/2013 4:35:10 PM]       Monitoring is turned on at 30 second intervals.
[12/9/2013 4:48:53 PM]       User clicked Pause Mining button for CGMiner with full API access. Attempting to stop...
[12/9/2013 4:48:53 PM]   [d] Miner.StopMining called: Miner=C:\gpu miner\cgminer.exe;ProcessID=1476 (True);BatchProcessID=0 (False);Config=C:\gpu miner\fst.conf;Arguments=--config "C:\gpu miner\fst.conf";HasFullAPIAccess=True;KillOnReadOnlyAccess=True;IsBatCmdFile=False;ProcessID=1476;.
[12/9/2013 4:48:56 PM]       CGMiner stopped successfully.
[12/9/2013 4:48:56 PM]       Mining paused successfully.
[12/9/2013 4:48:57 PM]   [d] 1 miner instances found, 1 ignored.
[12/9/2013 4:48:58 PM]   [d] Process 6928 ignored: Unable to match to a profile or communicate with on port 4028.
[12/9/2013 4:48:58 PM]       User clicked Start Mining button. Attempting to start miner...
[12/9/2013 4:48:58 PM]   [d] Miner.StartMining called. [Profile=NEW Fastcoin;MinerFile=C:\gpu miner\cgminer.exe;MinerFileUsed=C:\gpu miner\cgminer.exe;MinerExecutable=C:\gpu miner\cgminer.exe;Config=C:\gpu miner\temp.conf (C:\gpu miner\fst.conf);Arguments=--config "C:\gpu miner\temp.conf";Anticipated Port=4028;Used Port=4028;IsBatCmd=False;BatCmdPath=C:\gpu miner\cgminer.exe (C:\gpu miner\cgminer.exe);LastKnownInstanceCount=0/1]
[12/9/2013 4:48:59 PM]   [d] Miner.IsProcessAlive(8116) returns True. [First try, process is alive.] (CPID8116/CMPID=8116/CBPID=0)
[12/9/2013 4:48:59 PM]   [d] Waiting for CGMiner to initialize API. Process 8116 alive=True, minerid=8116.
[12/9/2013 4:49:00 PM]   [d] Watch for miner, expected process is 8116 (alive).
[12/9/2013 4:49:00 PM]   [d] CGMiner process is running using ProcessID 8116, expecting port 4028.
[12/9/2013 4:49:01 PM]   [d] CGMiner running but without API access... waiting to check again.
[12/9/2013 4:49:04 PM]   [d] CGMiner running but without API access... waiting to check again.
[12/9/2013 4:49:06 PM]   [d] CGMiner running and API initialized on port 4028, process matched to profile NEW Fastcoin.
[12/9/2013 4:49:06 PM]   [d] Finish wait for miner to start, process is 8116 (alive).
[12/9/2013 4:49:06 PM]       CGMiner started successfully.
[12/9/2013 4:50:00 PM]       User clicked Pause Mining button for CGMiner with full API access. Attempting to stop...
[12/9/2013 4:50:00 PM]   [d] Miner.StopMining called: Miner=C:\gpu miner\cgminer.exe;ProcessID=8116 (True);BatchProcessID=0 (False);Config=C:\gpu miner\temp.conf;Arguments=--config "C:\gpu miner\fst.conf";HasFullAPIAccess=True;KillOnReadOnlyAccess=True;IsBatCmdFile=False;ProcessID=8116;.
[12/9/2013 4:50:03 PM]       CGMiner stopped successfully.
[12/9/2013 4:50:03 PM]       Mining paused successfully.
[12/9/2013 4:50:03 PM]   [d] 1 miner instances found, 1 ignored.
[12/9/2013 4:50:04 PM]       User clicked Start Mining button. Attempting to start miner...
[12/9/2013 4:50:04 PM]   [d] Miner.IsProcessAlive(8116) returns False. [First try, test process is nothing.  ex: Process with an Id of 8116 is not running.] (CPID0/CMPID=0/CBPID=0)
[12/9/2013 4:50:04 PM]   [d] Miner.StartMining called. [Profile=NEW Fastcoin;MinerFile=C:\gpu miner\cgminer.exe;MinerFileUsed=C:\gpu miner\cgminer.exe;MinerExecutable=C:\gpu miner\cgminer.exe;Config=C:\gpu miner\temp.conf (C:\gpu miner\fst.conf);Arguments=--config "C:\gpu miner\temp.conf";Anticipated Port=4028;Used Port=4028;IsBatCmd=False;BatCmdPath=C:\gpu miner\cgminer.exe (C:\gpu miner\cgminer.exe);LastKnownInstanceCount=0/1]
[12/9/2013 4:50:05 PM]   [d] Miner.IsProcessAlive(5952) returns True. [First try, process is alive.] (CPID5952/CMPID=5952/CBPID=0)
[12/9/2013 4:50:05 PM]   [d] Waiting for CGMiner to initialize API. Process 5952 alive=True, minerid=5952.
[12/9/2013 4:50:06 PM]   [d] Watch for miner, expected process is 5952 (alive).
[12/9/2013 4:50:06 PM]   [d] CGMiner process is running using ProcessID 5952, expecting port 4028.
[12/9/2013 4:50:07 PM]   [d] CGMiner running but without API access... waiting to check again.
[12/9/2013 4:50:09 PM]   [d] CGMiner running but without API access... waiting to check again.
[12/9/2013 4:50:11 PM]   [d] CGMiner running and API initialized on port 4028, process matched to profile NEW Fastcoin.
[12/9/2013 4:50:11 PM]   [d] Finish wait for miner to start, process is 5952 (alive).
[12/9/2013 4:50:11 PM]       CGMiner started successfully.
[12/9/2013 4:50:12 PM]   [d] Coins.Refresh complete, 54 coin(s) present.
=================

Some of the ignored miner instances mentioned in the log may refer to the fact that I have cgminer-nogpu running some bit erupters in another task. That task has seemingly been unaffected by all this.

Any comments/suggestions from milone or anyone else?

+1

This is because we set the GPU clock and memory values into CGMiner... If you take thoes out, you'll notice that you'll no longer get the BSOD... I run into the same thing... This will be a GREAT stability enhancment to close out the CMD window instead of sending the "Quit" command...

Also, milone, man... You're amazing dude... This app is awesome...


I've ran into something over the past few day's and wanted to run it by you... So I've got 7 GPU's that I put into one profile... That works fine, but when one of those GPU's fail, the hash rate falls below the Scrypt hash rate but doesn't restart the miner... Any idea's why?

I'm running into the same issue on another miner I've got that's got 5 GPU's in it... Here's a screen shot of my settings (See that I've set the setting to 3.87)... I would appreciate if you could give me a suggestion for this feature to work?


http://i.imgur.com/NaqrsP6.png

Yes, I close cgminer clicking the X, always, I didn't even know there was a Quit command. I'll try it to see if I can replicate the crash.

EDIT: Yes, I replicated the bsod closing cgminer with its Quit command. I'm going to research this new angle.

EDIT 2: I've upgraded my cgminer from 3.2.1 to 3.7.2 to no avail. Still crashing.

Sorry, I  forgot to change the version number in the download link. It still pointed to 1.3.3.2, but the link text said 1.3.3.1. It should be fixed now.




How are you closing cgminer when you do it outside of CGWatcher? Clicking the X?  CGWatcher sends the 'quit' command when it stops the miner, which is different than clicking the X because the miner shuts down and returns clock and fan speeds back to their original values. This is the same as using the menu in cgminer to quit. Either way, I think I will make some changes to how CGWatcher performs restarts so there is no chance  it will restore clock and fan speeds just to have to bring them back up a few seconds later. That will be in the next update, and may provide better stability. But for stopping the miner, all CGWatcher does is send the 'quit' command to the miner and the miner closes itself after restoring clock speeds and cleaning up. When you click the X it never gets chance to do these things.

First of all thank you very mucn for this app, amazing stuff.

I've been using 1.3.3.2 version of CGWatcher a couple of days in 6 miners without problems but now I want to change the mined coin and half of my rigs crash when I try to stop or restart cgminer from inside cgwatcher. I can close cgminer manually without any problems. When I close cgminer from cgwatcher I get a BSOD related to the file ATIKMPAG.SYS every time. I've cleaned and updated AMD drivers without success. Any ideas?