Author

Topic: Ubuntu auto-start mining script issues... (Read 1713 times)

full member
Activity: 179
Merit: 100
June 29, 2011, 01:01:06 PM
#6
Well... in my case (ubuntu 10.10 64 bit) this sript now is working well:
Code:
#!/bin/bash
cd /home/user/
export LD_LIBRARY_PATH=/opt/ati-stream-sdk-v2.3-lnx64/lib/x86_64/:$LD_LIBRARY_PATH
AMDOverdriveCtrl -i 0 set.ovdr &
cd phoenix/
python phoenix.py -u http://worker:[email protected]:8332 -k phatk DEVICE=0  VECTORS BFI_INT FASTLOOP WORKSIZE=128 AGGRESSION=13 &
gnome-terminal -e 'python phoenix.py -u http://worker:[email protected]:8332 -k phatk DEVICE=0 VECTORS BFI_INT FASTLOOP WORKSIZE=128 AGGRESSION=5'
wait
Now no need to start it manually and if it stucks (I'll see it from btc-poolwatch.com page)  - just do 'killall phoenix.py & reebot' in SSH remote session snd one minute later both miners are up.
So you should find the right string with LD_LIBRARY_PATH  for your SDK version and 32b system -  insert it and relax.
newbie
Activity: 47
Merit: 0
Code:
Traceback (most recent call last):
... 
ImportError: libOpenCL.so.1: cannot open shared object file: No such file or directory

Can you try putting:
Code:
# or lib/x86 for 32bit
export LD_LIBRARY_PATH=/path/to/ati-stream-sdk/lib/x86_64

into your script
newbie
Activity: 27
Merit: 0
well, my ubuntu is 32bit and from what i read, things are little different within both versions...

please move this thread to the mining forum! thanks!
full member
Activity: 179
Merit: 100
got the same once - then pass again steps 15-17 from here and now every reboot all sytems are go without any mystical dances
newbie
Activity: 27
Merit: 0
*bump*
Well, it seems that nobody is able to help me here, would someone please put this thread ito the mining Forum, I'm sure soneone over there will know what I did wrong! Thanks!
newbie
Activity: 27
Merit: 0
Hey there,
I wrote a little script to start my miners which looks like this:

Code:
#!/bin/sh

while true
do
cd ~
cd phoenix-1.48/
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=11 DEVICE=0 WORKSIZE=256 &
sleep 2
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=5 DEVICE=0 WORKSIZE=256 &

sleep 5
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=11 DEVICE=1 WORKSIZE=256 &
sleep 2
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=5 DEVICE=1 WORKSIZE=256 &

sleep 5
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=11 DEVICE=2 WORKSIZE=256 &
sleep 2
./phoenix.py -u http://WORKERNAME:[email protected]:8332 -q 5 -k phatk VECTORS BFI_INT FASTLOOP AGGRESSION=5 DEVICE=2 WORKSIZE=256 &

sleep 7200
killall phoenix.py

done

The idea is, to start all three GPUs, with some delay to prevent a power surge. Although I am using two pools each, with different aggression to be on the safe site, if pools go down like they used to do last week. And because I'm still using phoenic 1.48, the script kills and restarts the whole process every two hours in case the miner crashes.
So far so good, this has been working fine for two weeks now.
Now I got a place to put up the rig in a Friends company office. (no heat and noise for me and although kWhs are much cheaper for businesses Wink )

To be make the rig absolutely stable, I now wanted to put my script into Ubuntus Startup applications, so that I can just call there and ask my friend to hit the reset button in case the system hangs up completely for whatever reason.

I put the following line into the startup-applications menu:
Code:
gnome-terminal -e '/bin/bash startup.sh'

But when I reset my system now, I get the following outut in the terminal that is supposed to run the script:
Code:
Traceback (most recent call last):
 
File "./phoenix.py", line 123, in
    miner.start(options)
File "/home/miner/phoenix-1.48/Miner.py", line 75, in start
    self.kernel = self.options.makeKernel(KernelInterface(self))
File "./phoenix.py", line 111, in makeKernel
    kernelModule = imp.load_module(module, file, filename, smt)
  File "kernels/phatk/__init__.py", line 23, in
import pyopencl as cl
File "/usr/local/lib/python2.7/dist-packages/pyopencl-0.92-py2.7-linux-i686.egg/pyopencl/__init__.py", line 3, in
import pyopencl._cl as _cl
ImportError: libOpenCL.so.1: cannot open shared object file: No such file or directory

Can Anyone tell me, why that is? I speculate it has to to with permissions (maybe the gnome-startup doesn't have the same permissions as I do, when i manually call the script?

Thank you guys!
Jump to: