BAMT version 0.5 - Easy USB based mining Linux with farm wide management tools - page 36.

david113

member

Activity: 63

Merit: 10

Quote from: Joshwaa on May 25, 2012, 10:30:56 AM

Yes please donate to the 7970 & 64bit BAMT fund!!!!

How much BTC´s left to support 7970s??

abracadabra

hero member

Activity: 956

Merit: 1001

Ok... here's what i've put together with some help from some friends (thx C)

create a check_network.sh in /etc/bamt this script will check for internet connection. if there is none, it will bring the ethernet interface down, wait a bit, then bring it back up.

Code:

#! /bin/sh

ping -c1 -W5 google.com > /dev/null  || (
    echo "down"
    ifdown eth0
    sleep 5
    ifup eth0
)

create a file called crontab in /etc/bamt - this will be a file used to create a crontab entry that will run the check_network.sh script every minute.

Code:

* * * * * /etc/bamt/check_network.sh

edit /etc/rc.local to add the following line just before the "exit 0" - upon bootup this will add a crontab entry.

Code:

crontab /etc/bamt/crontab

at this point you can reboot, or simply type "crontab /etc/bamt/crontab"

I've been running rigs for over 2 days straight now that would have normally lost connection after only 12 hours.

abracadabra

hero member

Activity: 956

Merit: 1001

I believe I've been able to find a solution to the rigs losing network connection. I had a few rigs that would disconnect/drop network connection after 6-12 hours. With my fix, the rigs have been running for a bit over 14 hours and they are still running. I want to test for a couple more days and then I'll post something up in this thread.

Removing network-manager did not solve the problem for me. Neither was setting a static IP and neither the "fix" detailed here: http://www.bitcointrading.com/forum/index.php?PHPSESSID=528oms2jquas75afn1t6fp45n5&topic=416.msg1235#msg1235

This is not a problem due to the rig hardware. I have exact duplicate rigs deployed in multiple locations. I believe it has something to do with the router/dhcp server. I only have this problem when I bring the rigs home to troubleshoot a gpu. When the rigs are deployed offsite to networks with different network hardware, the loss of network connection does not occur.

bulanula

hero member

Activity: 518

Merit: 500

Quote from: Transisto on May 27, 2012, 11:01:44 AM

Debian connection lost problem,

Is there a bug report for this on debian's forums ?

I would say this has happened / happening on 2/5 of my rigs,

Quote from: lodcrapo

It basically just boils down the network-manager package. This software has a bug that seems to effect some systems, just google around and you'll find lots of ubuntu users with the exact same issue. I didn't find any clear explanations or what the proper fix is in a quick search. It does seem to be limited to certain network cards.

rather than removing the network manager entirely (as suggested by the external link), if someone finds an actual fix that would be preferable.
removing the network manager will remove the issue, but it will take away your ability to configure wireless networks (and wired) using the easy gui, so you'll have to edit the network config files by hand. not a big deal if you know how to do that, but not an ideal "fix" in any case.

Remove the network-manager.

It is dead simple to change IP config in Debian using CLI.

Not worth having this damn bug IMHO

Transisto

donator

Activity: 1731

Merit: 1008

Debian connection lost problem,

Is there a bug report for this on debian's forums ?

I would say this has happened / happening on 2/5 of my rigs,

Quote from: lodcrapo

It basically just boils down the network-manager package. This software has a bug that seems to effect some systems, just google around and you'll find lots of ubuntu users with the exact same issue. I didn't find any clear explanations or what the proper fix is in a quick search. It does seem to be limited to certain network cards.

rather than removing the network manager entirely (as suggested by the external link), if someone finds an actual fix that would be preferable.
removing the network manager will remove the issue, but it will take away your ability to configure wireless networks (and wired) using the easy gui, so you'll have to edit the network config files by hand. not a big deal if you know how to do that, but not an ideal "fix" in any case.

Joshwaa

hero member

Activity: 497

Merit: 500

Yes please donate to the 7970 & 64bit BAMT fund!!!!

abracadabra

hero member

Activity: 956

Merit: 1001

Quote from: tossil on May 24, 2012, 11:10:26 PM

I've run this on two rigs, one with 3x 7970's and another 2x 7970's. On both rigs, after I set the bamt.conf and pool I can only get the first card to mine with cgminer. If I run it by command line, I can get all cards to mine.

I could be wrong, but bamt does not support 7970s. Please read previous posts about encouraging lodcrappo to develop BAMT for 7970s if you want. There's a "donation" fund. Once it hits the goal, he'll get on it. Wink

abracadabra

hero member

Activity: 956

Merit: 1001

Quote from: viperzero on May 24, 2012, 04:40:25 AM

Quote from: jamesg on May 23, 2012, 07:36:15 AM

Most 6xxx series cards you'll need to flash the bios to get low memory with overclocking.

Thankyou for response. That's what I was afraid of. So I guess have to dig out some firmware & flashing tool and take card out to another (windows)machine for flashing. I guess there is no flashing utility for Debian either? I would appreciate if anyone has sapphire 6870 OC (core 900 mem 1150) underclocked firmware. I guess 900/300 would be good clocks for this. Hopefully changing to FPGA mining soon & wouldn't like to make too much work If I sell my GPU cards anyway soon.

I am able to underclock memory to -125 from core (ie. core 900 mem 775) without having to mess with the firmware. But if you are looking for a greater than 125 difference core to mem then you'll have to go the firmware route.

tossil

newbie

Activity: 16

Merit: 0

I've run this on two rigs, one with 3x 7970's and another 2x 7970's. On both rigs, after I set the bamt.conf and pool I can only get the first card to mine with cgminer. If I run it by command line, I can get all cards to mine.

When I hit 1 for GPU1 while in GPUMON I get this... (hitting 2 for GPU2 is the same)

cgminer version 2.4.1 - Started: [2012-05-25 04:08:20]
--------------------------------------------------------------------------------

TQ: 2 ST: 2 SS: 0 DW: 0 NB: 1 LW: 2 GF: 0 RF: 0
Connected to http://us2.eclipsemc.com:8337 with LP as user me
Block: 000007274358314b696a8802185423fa... Started: [04:08:20]
--------------------------------------------------------------------------------
[P]ool management [G]PU management [ S]ettings [D]isplay options [Q]uit
GPU 0: 60.0C 3411RPM | 0.0/ 0.0Mh/s | A:1 R:0 HW:0 U:600.00/m I:11
--------------------------------------------------------------------------------
[2012-05-25 04:08:19] Started cgminer 2.4.1
[2012-05-25 04:08:19] Started cgminer 2.4.1
[2012-05-25 04:08:20] ADL found more devices than opencl!
[2012-05-25 04:08:20] There is possibly at least one GPU that doesn't support OpenCL
[2012-05-25 04:08:20] Use the gpu map feature to reliably map OpenCL to ADL
[2012-05-25 04:08:20] WARNING: Number of OpenCL and ADL devices did not match!
[2012-05-25 04:08:20] Hardware monitoring may NOT match up with devices!
[2012-05-25 04:08:20] Probing for an alive pool
[2012-05-25 04:08:20] Long-polling activated for http://us2.eclipsemc.com:8337/LP
[2012-05-25 04:08:20] Pool 0 http://us2.eclipsemc.com:8337 alive
[2012-05-25 04:08:20] Accepted cd868403.d19db6ce GPU 0
[2012-05-25 04:08:25] API bind to port 4028 failed - trying again in 30sec

If I quit out of this and run it by command line, I can get all three to run.

Here is my command line I'm running: /opt/miners/cgminer/cgminer -o http://us2.eclipsemc.com:8337 -u me -p pass --intensity 11 --api-listen --api-port 4028

Does anyone have any idea where I should I start looking?

Thanks in advance for any advice,

-Tossil

EDITED to Add...

I ran CGMINER with -n and I see there is an error -1.

root@bamt-miner:/home/user# /opt/miners/cgminer/cgminer -o http://us2.eclipsemc.com:8337 -u me -p pass --intensity 11 --api-listen --api-port 4028 -n
[2012-05-25 12:49:58] CL Platform 0 vendor: Advanced Micro Devices, Inc.
[2012-05-25 12:49:58] CL Platform 0 name: AMD Accelerated Parallel Processing
[2012-05-25 12:49:58] CL Platform 0 version: OpenCL 1.1 AMD-APP (844.4)
[2012-05-25 12:49:58] Platform 0 devices: 3
[2012-05-25 12:49:58] 0 Tahiti
[2012-05-25 12:49:58] 1 Tahiti
[2012-05-25 12:49:58] 2 Tahiti
[2012-05-25 12:49:58] ADL Initialisation Error! Error -1!
[2012-05-25 12:49:58] 3 GPU devices max detected
root@bamt-miner:/home/user#

BitMinerN8

hero member

Activity: 626

Merit: 500

Mining since May 2011.

Quote from: TheHarbinger on May 24, 2012, 05:40:57 PM

Quote from: BitMinerN8 on May 24, 2012, 05:21:12 PM

Quote from: BitMinerN8 on May 24, 2012, 04:36:16 PM

Quote from: Inaba on May 24, 2012, 03:46:08 PM

You need to use -S /dev/ttyUSBx for each additional past the first.

Thanks, that worked. A tip your way too, later this evening.

I think I'm 95% there. Grin

And lastly, does anyone know of a way to get the GPU's and BFL's to get reported to the BAMT Web Stat page? (munin) Currently it only shows the GPU's. I tried:

Code:

do_cgsnoop: 1

But that did not help.

Don't believe you can without rewriting munin yourself. There are no monitoring tools for FPGA systems built into BAMT, and lodcrappo has stated that he has no plans to include them.

Ahhh, so I guess I'll go back to Win7 on the BFL rig and then use cgsnoop on a BAMT rig to get it into mgpumon.

TheHarbinger

sr. member

Activity: 378

Merit: 250

Why is it so damn hot in here?

Quote from: BitMinerN8 on May 24, 2012, 05:21:12 PM

Quote from: BitMinerN8 on May 24, 2012, 04:36:16 PM

Quote from: Inaba on May 24, 2012, 03:46:08 PM

You need to use -S /dev/ttyUSBx for each additional past the first.

Thanks, that worked. A tip your way too, later this evening.

I think I'm 95% there. Grin

And lastly, does anyone know of a way to get the GPU's and BFL's to get reported to the BAMT Web Stat page? (munin) Currently it only shows the GPU's. I tried:

Code:

do_cgsnoop: 1

But that did not help.

Don't believe you can without rewriting munin yourself. There are no monitoring tools for FPGA systems built into BAMT, and lodcrappo has stated that he has no plans to include them.

BitMinerN8

hero member

Activity: 626

Merit: 500

Mining since May 2011.

Quote from: BitMinerN8 on May 24, 2012, 04:36:16 PM

Quote from: Inaba on May 24, 2012, 03:46:08 PM

You need to use -S /dev/ttyUSBx for each additional past the first.

Thanks, that worked. A tip your way too, later this evening.

I think I'm 95% there. Grin

And lastly, does anyone know of a way to get the GPU's and BFL's to get reported to the BAMT Web Stat page? (munin) Currently it only shows the GPU's. I tried:

Code:

do_cgsnoop: 1

But that did not help.

BitMinerN8

hero member

Activity: 626

Merit: 500

Mining since May 2011.

Quote from: Inaba on May 24, 2012, 03:46:08 PM

You need to use -S /dev/ttyUSBx for each additional past the first.

Thanks, that worked. A tip your way too, later this evening.

Inaba

legendary

Activity: 1260

Merit: 1000

You need to use -S /dev/ttyUSBx for each additional past the first.

BitMinerN8

hero member

Activity: 626

Merit: 500

Mining since May 2011.

Quote from: TheHarbinger on May 24, 2012, 02:50:37 PM

Quote from: BitMinerN8 on May 24, 2012, 02:30:04 PM

Looking for a little guidance, I was trying to get a few BFL's working on BAMT, but I seem to be missing something:

I am assuming that I need to run CGMINER to get it working so I added the following to my bamt.conf:

Code:

cgminer: 1
cgminer_opts: --api-listen -o pool:port -u user -p pass -I 9 (and other pools, options, etc.)
cgminer_port: 4028

And in every GPU section I also added a new line:

Code:

GPU0
  cgminer: 1

After restarting mining, it seems to work for GPU's, and I am submitting shares but I don't see the CGMINER stats screen anywhere. How do you access that? (I am using SSH to admin the rig, did not hook up a monitor to check on the console directly)

Also where do the BFL's show up? (linux noob) On Windows 7 I just added the -S bitforce:\\.\COM12 etc for each unit. Does anyone know where to put that info? I have it all working under Windows 7, CGMINER shows GPU's and BFL's, I just want to get back on BAMT for obvious reasons.

From root, either through ssh or local;

Code:

screen -ls

That will show you what screen number cgminer is running on.
Then ;

Code:

screen -r ####

Where the #### is the number returned by "screen -ls"

When you are done, you can't just close the terminal window, that will kill the cgminer process, you have to "detach" from the screen.

Control + a, d

Cool, that worked to connect. It seems to have auto added 1 BFL. Any tips on how to find or add the other 4? (I have 5 total.)

Thanks! I'll send you a donation when I get to my wallet.

TheHarbinger

sr. member

Activity: 378

Merit: 250

Why is it so damn hot in here?

Quote from: BitMinerN8 on May 24, 2012, 02:30:04 PM

Looking for a little guidance, I was trying to get a few BFL's working on BAMT, but I seem to be missing something:

I am assuming that I need to run CGMINER to get it working so I added the following to my bamt.conf:

Code:

cgminer: 1
cgminer_opts: --api-listen -o pool:port -u user -p pass -I 9 (and other pools, options, etc.)
cgminer_port: 4028

And in every GPU section I also added a new line:

Code:

GPU0
  cgminer: 1

After restarting mining, it seems to work for GPU's, and I am submitting shares but I don't see the CGMINER stats screen anywhere. How do you access that? (I am using SSH to admin the rig, did not hook up a monitor to check on the console directly)

Also where do the BFL's show up? (linux noob) On Windows 7 I just added the -S bitforce:\\.\COM12 etc for each unit. Does anyone know where to put that info? I have it all working under Windows 7, CGMINER shows GPU's and BFL's, I just want to get back on BAMT for obvious reasons.

From root, either through ssh or local;

Code:

screen -ls

That will show you what screen number cgminer is running on.
Then ;

Code:

screen -r ####

Where the #### is the number returned by "screen -ls"

When you are done, you can't just close the terminal window, that will kill the cgminer process, you have to "detach" from the screen.

Control + a, d

BitMinerN8

hero member

Activity: 626

Merit: 500

Mining since May 2011.

Looking for a little guidance, I was trying to get a few BFL's working on BAMT, but I seem to be missing something:

I am assuming that I need to run CGMINER to get it working so I added the following to my bamt.conf:

Code:

cgminer: 1
cgminer_opts: --api-listen -o pool:port -u user -p pass -I 9 (and other pools, options, etc.)
cgminer_port: 4028

And in every GPU section I also added a new line:

Code:

GPU0
  cgminer: 1

After restarting mining, it seems to work for GPU's, and I am submitting shares but I don't see the CGMINER stats screen anywhere. How do you access that? (I am using SSH to admin the rig, did not hook up a monitor to check on the console directly)

Also where do the BFL's show up? (linux noob) On Windows 7 I just added the -S bitforce:\\.\COM12 etc for each unit. Does anyone know where to put that info? I have it all working under Windows 7, CGMINER shows GPU's and BFL's, I just want to get back on BAMT for obvious reasons.

viperzero

member

Activity: 301

Merit: 10

Quote from: jamesg on May 23, 2012, 07:36:15 AM

Most 6xxx series cards you'll need to flash the bios to get low memory with overclocking.

Thankyou for response. That's what I was afraid of. So I guess have to dig out some firmware & flashing tool and take card out to another (windows)machine for flashing. I guess there is no flashing utility for Debian either? I would appreciate if anyone has sapphire 6870 OC (core 900 mem 1150) underclocked firmware. I guess 900/300 would be good clocks for this. Hopefully changing to FPGA mining soon & wouldn't like to make too much work If I sell my GPU cards anyway soon.

jamesg

vip

Activity: 1358

Merit: 1000

AKA: gigavps

Quote from: viperzero on May 23, 2012, 07:32:58 AM

Hi,

Bamt seem to be an amazing cp for mining rigs, tyvm sending some coins after get this problem solved:

I have Gigabyte 990fx mobo, got it booting thanks to additional image & second reader. Originally rig had 4 x 5850 but I had to replace 1 with Sapphire Radeon 6870 OC edition.

I can downclock core & memory easily on 5850s but not on 6870 OC. I think log revealed the problem:

(--) PCI: (0:2:0:0) 1002:6738:174b:e178 ATI Technologies Inc rev 0, Mem @ 0xc0000000/268435456, 0xfdec0000/131072, I/O $ @ 0x0000de00/256, BIOS @ 0x? Huh

?/131072

So seems 6870 bios (starting address?) location is missing. Does this mean that bios is locked or just not recognized right?
Would really appreciate help because this rotten egg in rig totally messes up both power consumption & heat values.

Thankyou anyway for BAMT & all who have been contributing.

Most 6xxx series cards you'll need to flash the bios to get low memory with overclocking.

viperzero

member

Activity: 301

Merit: 10

Hi,

Bamt seem to be an amazing cp for mining rigs, tyvm sending some coins after get this problem solved:

I have Gigabyte 990fx mobo, got it booting thanks to additional image & second reader. Originally rig had 4 x 5850 but I had to replace 1 with Sapphire Radeon 6870 OC edition.

I can downclock core & memory easily on 5850s but not on 6870 OC. I think log revealed the problem:

(--) PCI: (0:2:0:0) 1002:6738:174b:e178 ATI Technologies Inc rev 0, Mem @ 0xc0000000/268435456, 0xfdec0000/131072, I/O $ @ 0x0000de00/256, BIOS @ 0x? Huh

?/131072

So seems 6870 bios (starting address?) location is missing. Does this mean that bios is locked or just not recognized right?
Would really appreciate help because this rotten egg in rig totally messes up both power consumption & heat values.

Thankyou anyway for BAMT & all who have been contributing.

Topic: BAMT version 0.5 - Easy USB based mining Linux with farm wide management tools - page 36. (Read 324185 times)