Topic: Cointerra Hardware Support **Unofficial - page 7. (Read 56948 times)

http://en.wikipedia.org/wiki/Udhcpc

Hello,

Don't see this in any of my beaglebones, when I do the top command in one I find the below, does anyone know what this is?

180     1 root     S     1136   0%   0% udhcpc -R -b -x hostname TerraMiner -t 20 -p /var/run/udhcpc.eth0.pid -i eth0

Thank You, there was a problem with my pool setup.  strange as exactly same thing in other terraminers but they work.  I am using stratum proxy server and had everything pointed to stratum+tcp://x.x.x.x:yyyy and changed to http://x.x.x.x:yyyy.

now they work.  Thank you.

On a fresh reboot, what is in /tmp/cgminer.log ?

no, there is no unresponsive cgminer, when I check with the top command, there is no cgminer running. 

I will even click on the restart cgminer in the web gui and it still does not show up.  So I manually start the ./cgminer.  If I do not kill the crond listed in top, the manual ./cgminer instance I start will get killed.

So, is there an unresponsive cgminer process running? If so, you could try getting that process to restart with something like

sudo kill -HUP processIDofCgminer

edit: might be -SIGHUP or just -1 on some versions.

Yes, it when I log out it kills the process. 

The problem I am having is some of my beaglebones will not start cgminer (not from reboot or web gui confi interface).  the crond is running and it tries to restart.  so I ssh in, go to opt directory and manually start ./cgminer.  I have to kill crond or it will kill my instance of cgminer.  when I close the connection it stops cgminer.

In standard linux, if you put an ampersand after the command it unties it from the console you ran it from, don't know it if unties it from the user though, i.e. log out and process dies.

In some linux systems when you ssh in, you can type "screen" to start a process running, then Ctrl A then D to keep the process running and end your ssh connection.  (if you end your ssh connection you will end the process)

The screen command does not work with the beagle bone linux software.  does anyone know how to do this type command for the linux on the beagle bone?

Hello all,

First running
Terraminer IV
CGMINER 4.3.5
FIRMWARE 0.7.56
Some cores are off. Trying to troubleshoot, board issue, power issue, end of the world as we know it issue, etc....

Is there anyplace that has the definitions for the fields in the STATS output in CGMINER 4.3.5?
ADVANCED-->STATS

Example fields;
 CorePerfMode0
 CoreFmatch0

I've looked and I can't find any manual or instruction. I'm looking for field definitions and system relations. I can guess on the obvious. But how they relate or specific conditions I may be unaware of.


Thanks

Thanks.  maybe someone just starts a troubleshooting guide in excel & save it to google docs.  If no one else starts it I will give it a try.

Regarding my USB errors all the lights are good (no red or out) and do the night rider sequence.

Here's what I know about the D45 light:

Obviously if it is on and red that's bad!

I had a red D45 on one board that was fixed by simply re-applying the thermal paste - this can be verified by downloading the diagnostic file and following the logic - i.e. everything boots, then slowly overheats and shuts down - as opposed to not starting at all.  So, that was my first instance with the red D45

Furthermore, I know if a pump is not working or not connected it will go straight to the red D45 without even trying to start beyond the first flicker of all the lights D37-D44.

What I did (I'm not smart with programming) is download NotePad++ and used it to search and highlight words like "overtemp", "fail", and "error" in the diagnostic file to see where things are going wrong.  Although there is a lot I do not understand - I simply compared it to one of my normal machines.

By the way, SW1 (push button) on the back of the board is apparently the reset switch - used it to diagnose things while it was running - simply restarts the respective board.

My assumption is: D37-D44 are lights representing each of the 8 cores and by sequencing yellow it means everything is normal.  If it is skipped in the sequence it must be bad or not responding.  D45 is some sort of error indicator whereas OFF is normal and Red is bad but does not necessarily always mean the same thing?

Would be nice if between all of us here we could put together a guide of known issues/fixes and a decoding key for the diagnostic report in a public document that doesn't require reading this entire thread - anyone know how to do this?

More to follow as I troubleshoot...

USB Errors.

Does anyone have any guesses to try to fix?  screenshot is with a direct connect to the board USB.

it happens with beaglebone or usb direct.

it continually connects then disconnects.  Any ideas to try?  I have 4 or 5 boards like this so it would be nice to find a fix.  Thanks

Wrote them in support.
Support this pathetic parody of specialists. But sent the link.
The first time it was a link to the damaged file with the image file.
The second time the file is normal, but apparently the image itself is damaged.  https://yadi.sk/d/y-Q05244dcTeP
I tried to install the firmware image, but I did not succeed. http://youtu.be/P9czTTtjHB0    http://youtu.be/g3V-LgOqTmo
At the same time updates from the BeagleBoneBlack installed without any problems at all.

Look. Maybe someone can.
Describe the process then.

well from what i've learned over the past few days if you open up the case of the miner you will see a knight rider type of light on the board. if not then there could be a problem. if you only want to run one side at a time make sure everything is plugged into that same power supply. i had to move my power for the beagle bone to the opposite power supply and unplug the dead boards usb cable to get mine hashing

Hey Folks,

Just "upgraded" to a broken terraminer. One board not working (PSU side). Price was good, wanted to be running some "network support" hash, BFLs are past it. Speed read the entire thread, seeing a bunch of different possible problems. Previous owner tried swapping PSUs over, bad side PSU ran good board, but that's all he tried. He thought it might be a dead pump. No crispy critters turned up by eyeball so far.

So, is there a good order to test/try things in?

Also not 100% sure at the moment about whether I have the power in one location to run both sides at once. Anyone decased a board completely, run independent? Saw the post about running on a laptop, that's half of it I guess.

Another thing is that I was wondering if anyone had posted up approx hash/power/efficiency for each of the power steps available. Thinking I may have an easier time with power hookup if I can run one side at 900W ish, the other full bore.

Haven't tried powering it up yet, need to get organised, just wondering if I should have it apart for inspection first or, try powering both to see what happens.

Thanks for the hints, PS total Cointerra n00b, didn't really follow developments in past.

yepp its pretty much going to end up in the trash lol. that was a big waste of money. only one board is currently working. the other one wont even boot up. if you have to go through this much bullshit ? its not even worth it. if anyone has any ideas why one of the boards wont boot let me know. otherwise it'll be stripped and sold.


the board was working.after the thermal paste but still had the red light and the miner shut down unexpectadly. when i tried to reboot the board would not start again. and i had to unplug the usb from it

EDIT: the board that works. works amazing at low temperatures. i wonder how long it will last outside.

miningrigrentals may have the same issue as nicehash.  I have the same issue.  Nicehash says to setup a stratum proxy server with exranounce subscription.  I am working on setting one up so I can help on that.

inferno:  regarding heat issues, even after adding new thermal paste, the glycol will drop out when they are hot and deposit on those copper plates.  You will have to take it apart and clean it.  and refill it making sure no air is in there.  There is barely any fluid volume in there so even a small air bubble may cause a problem.

here is the full report after re thermal pasting. one of the boards has one light that doesnt light up at all. and one that is red. in that series of led's d42 does not light and d45 is red on the board beside the power supplies.
http://pastebin.com/PnuYUhLz

currently the power stepping is at 7. any higher and both the board throttles back down to this freq. same as before. when i took the cooler off. they were so coated in thermal paste it was ridiculous. the moats for the asic dies were filled in even. all i did was replace with a dab on the top of each die. like a normal CPU. so if im wrong here and the moats need to be filled to let me know. the only thing i didnt do was take off the copper block from the pump. since i noticed a few people have done that with better results.

Also i noticed on the boards there are 2, 2 pin connectors. labeled jp1 and jp2. the board furthest away from the power supply has a connector on jp1 connecting the 2 pins together. but the rest of the boards do not

EDIT: ive tried every firmware possible. now the board with the previous red light. lights all of the leds up clicks and shuts down. the board that is working. has a red light on all firmwares below 0.76. so now i've got half a miner. that i cant fully power step

For who want to help us, here is an useful briefing of how to extract the eMMC image for BeagleBone Black controller of TerraMiner IV unit.

http://elinux.org/BeagleBone_Black_Extracting_eMMC_contents


Thanks to all for any help! ;-)