Topic: LinuxCoin A lightweight Debian based OS with everything ready to go. - page 63. (Read 285117 times)

Further notes on persistence and automatic startup.

I found it much easier to boot the CD, plug the drive in, and partition it there.  You can also set up the ext4 filesystem at this step.  Set the type of partition 1 to "C".  Then move it back to the windows box to format.

If you are using unetbootin, save yourself a lot of hassle by opening syslinux.cfg with WordPad and adding "persistent" to the end of the first append line (under label unetbootindefault).

The autostart is in /home/user/.config/autostart/auto.desktop


Keep in mind that this won't run as root, so you'll have to diddle the permissions and/or use setuid on the shell script that it runs.

A couple of my miners are 50 miles away from my home, so I just finished setting all of this up, along with a watchdog to reboot the boxes as needed.  I figured since I had to figure a bunch of this out the hard way, I'd write this post while it was still fresh in my mind.  Hope it helps someone.

BTW, I then received -


after - IOError: [Errno 13] Permission denied: '58ad084353512b9677717d55606d6f91.elf'

31/05/2011 06:46:38, long poll: new block 000002a137bc91ed 


Does this mean I received work but it expired?

I tried "start_mining" with poclbm.... but I get -


Exception in thread Thread-2:
Traceback (most recent call last):
  File "/usr/lib/python2.6/threading.py", line 532, in __bootstrap_inner
    self.run()
  File "/usr/lib/python2.6/threading.py", line 484, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/opt/miners/poclbm/BitcoinMiner.py", line 272, in miningThread
    self.loadKernel()
  File "/opt/miners/poclbm/BitcoinMiner.py", line 373, in loadKernel
    binaryW = open(cacheName, 'wb')
IOError: [Errno 13] Permission denied: '58ad084353512b9677717d55606d6f91.elf'



I saw someone else have this same issue..... but there didn't appears to be a resolution.

Any ideas?

Thanks.

Thx for your suggestion. I just attempted phoenix on one card. I get a similar result. It seems to alternate between 80 & 120MH/s instead of 275MH/s

Is there an easy way to increase the maximum clock rate limit in Overdrive?  I can only seem to increase the clocks by 50 mhz (ATI 5870).  Is RBE the only option?

yes a hard disk install would be good... this seems very unstable, 1 GPU seems to stop working after a while on each rig, still visible with aticonfig, but when trying to start it up again ends in a kernel panic, freeze up and need for a reboot.... annoying!! maybe its just my motherboards :/


edit: maybe running the processes with "screen" or in the background with "&" was giving the kernel panics?  ive opened a terminal window for each process and minimised them all, now seems to be stable for the last hour

edit2: no difference.... maybe ill try phoenix miner instead

edit3: phoenix with poclbm kernel is slower than poclmb on my 6990s ....  now i have uptime of 12 hours+ on one rig, i just used the ROOT terminal for each instance of poclbm and minimised each window... lets see how long this lasts.

Please make a regular install version.  Not everyone wants to run off a live CD / flash drive.  I prefer to run off a spare laptop HD (2.5W).

I tried to boot off of this yesterday and it was just an orgy of errors.

may I add this was while using the start_miner script..

i have now killed the syslogd process, installed screen and started all miners from command line within their own instance of screen..  looking good so far

i can not seem to get a good uptime with this.. either running it from dvd drive or from usb flash drive

after 12-24 hours i get error in the console from my ssh session


This happens on both my mining rigs, have tried persistant and normal boot

Try pheonix or diablo miner

Hi,

drgr33n, thanks for this amazing package. Once I'll get some coins in my vault, I'll send you a thank you.

I've been reading about mining for about a month now. I finally received all my parts and finished the setup yesterday.

I was hoping if someone can give me some pointers on how to improve my current 3 GPU setup (individual process, no crossfire) that is not producing properly:
GPU1: 275 MH/s
GPU2: 1 to 275 MH/s
GPU3: 1 to 275 MH/s

I have a default LinuxCoin setup and mining with 3 processes of poclbm (I didn't use the start_mining command):
python poclbm.py --user=xxxx.xxx1 --pass=xxxx -o mining.bitcoin.cz -p 8332 -d0 -v -w128 -f 1 --verbose
python poclbm.py --user=xxxx.xxx2 --pass=xxxx -o mining.bitcoin.cz -p 8332 -d1 -v -w128 -f 1 --verbose
python poclbm.py --user=xxxx.xxx3 --pass=xxxx -o mining.bitcoin.cz -p 8332 -d2 -v -w128 -f 1 --verbose

Hardware:
MSI 890FXA-GD70 - http://www.msi.com/product/mb/890FXA-GD70.html
3 x 5850 - http://xfxforce.com/en-us/products/GraphicCards/HD%205000series/5850.aspx
AMD Sempron 145
2GB DDR3
Corsair 1200W power supply

For some reason:
- when GPU1 is running, GPU2 & GPU3 produce a very variable output, ranging anywhere between 1 to 275MH/s.
- when I stop GPU1, GPU2 & GPU3 produce 275MH/s each.
- I start GPU1, GPU2 & GPU3 are back to low production numbers.

I would also like to note that I have not yet done any special overclocking adjustments at any level.

Please find attached a screenshot as well as a picture of the current hardware setup.

Any advice/help/tips is appreciated.

Thank you

Higher Res
http://www.filefactory.com/file/cb8e098/n/MiningRig.JPG
http://www.filefactory.com/file/cb8e1ab/n/LinuxCoin_3GPUs.jpg

P.S. Upload folder on the forum was full, I was not able to attach an image.

Only device 0 gets auto-loaded by the gui. You'll need to use the commandline to load up another gui for other gpus.

details: http://sourceforge.net/projects/amdovdrvctrl/forums/forum/1335753/topic/4424733

....and in short:
1. find the device index using the help switch:

2. check the index numbers of the active devices, will be 0 and some other integer (in my case, 0 and 5).

3. then:

This will load up a separate gui for the card you want.

eg. for me:

More cards? Wash, rinse, repeat.

I actually got Linuxcoin to run after like 10 tries. However, I could only overclock the first GPU using the AMDOverDriveCtrl provided. Can someone tell me how I could OC the other cards I have?

It is marginal in fact. Its a 1.2kw power supply but that is output power, it was taking 1.3kw on input. You would expect the input power to be more than the output power because it is not 100% efficient.

Temperatures were high, mid 80's but it would have more cooling eventually.

Its also likely that some memory under clocking would not harm hash rates and would reduce consumption.

I Look on the bright side, it didn't blow with a large bang.

Sounds to me more like your power supply is not capable of providing enough current
What were the temperatures?

I am well impressed with this.

I had ordered a rig with 3 6990 cards in and was putting it together. I had really come to the conclusion that nothing worked as expected with ubuntu 10.4,10.10 or 11.04, sdk 2.1,2.3, 2.4 and catalyst 11.5, because the installer on 11.3 seems faulty with certain versions of Ubuntu. I have spent several days trying to fix this. Yes, many people say blithly use sdk 2.4, they will work, but they are using catalyst 11.3 which doesnt even install... Aaagh.

This evening dramatically I had all 6 cores hashing, and the hot air was blowing round my feet like a furnace, and the Kwh meter said 1.3 Kw, and I was scp' ing  my temeperature monitor to monitor what the core temps of the cards were. Suddenly, everything suddenly went quit. There were no spinning fans any more.  Power supply had got too hot and shut down.

Of course, this is the moment where you really would rather not have a volatile file system!

Anyway, perhaps I need to remove 1 card and let the power supply breathe...

Hey Valentin

Thanks for your post. One of the best I've seen laid out lol


This seems to stem from the same issue with the file manager. If you click the icon a few times LXDE crashes and resets. I'm looking into this but have been very busy over the last few days with work.


Also on the todo list. I'm thinking I might use zenity to create a little GUI for the final version ?.

There are several ways you can do this. You can add a line into /etc/rc.local just before the exit. Or if you want the WDM to load then run the commands you can follow the guide in the link provided.

http://linux.koolsolutions.com/2009/09/01/howto-auto-launch-programsscripts-on-lxde-startup/

Hope this helps.

Hi!

Your product is great. Not sure if anyone answered this, but you can use AMDOverdriveCtrl using flag -i=0 -i =1 etc for each adapter and it works.
I use it like this: /usr/bin/AMDOverdriveCtrl -i=0 Profilename.ovdr

The profilename you can save, but it has to be in the same folder (couldn't get it to work with paths) so I copy it in /usr/bin everytime.

Kill screensaver (takes resources when it's on): killall xscreensaver
Please disable screensaver by default in next version.

Issues:
- file manager (known)
- shutdown not working (known)
- sometimes (after moving the mouse, not sure exactly what is going on) the screen shows stars and login (where I just hit Enter), and everything that was running is gone -- it's like a soft reset (it doesn't reload the kernel, but everything is gone, maybe I have a memory problem, I don't know);
- start_mining is not useful (too complicated, I'd rather use manual commands)

Big question:

How do I run a start-up script? Everytime I reboot I have to manually run a script, which kills screensaver, runs AMDOverdriveCtrl in batch (using -b flag everything works except fan control), sets fans (using aticonfig), runs clients. I would love to be able to have this started up manually, so I just hit the reset button and it's good to go.

Thanks!

Valentin

Can you give me more info about hardware etc. Unable to connect to RPC normally means your internet connection is sketchy.