I just spent some time tweaking my shell script that I use to manage my miner instances.
* Central control of all miners - lets you switch pools on demand - the address of the pool you connect to is fetched from your web server
* Email alerts - yummy spam!
* Overheat protection - can shutdown miner at preset threshold
* Idle Miner protection - can restart stick miners
You can get it here: https://github.com/jsidhu/Bitcoin-Miner-Script
Here's the Readme:
This is a script that manages your bitcoin miners.
This will start phoenix in a detached screen session, the screen sesion name is set to the MINER_ID-Miner as specified in the settings.
For phoenix miner to work in a detached screen session, you should edit your ~/.screenrc and set the appropriate environment variables, I have the following in my ~/.screenrc
setenv LD_LIBRARY_PATH "/opt/AMD-APP-SDK-v2.4-lnx64/lib/x86_64/:/opt/ati-stream-sdk-v2.1-lnx64/lib/x86_64/"
setenv DISPLAY ":0"
This script uses a file hosted on your web server to figure out where to "point" your miner to. It also checks to make sure that the Temperature of your GPU is under a set limit and that the LOAD on your GPU is above a specified threshold.
IT requires that you have wget and sendemail installed. To install both of these packages, issue a apt-get install wget sendemail
* please notice that its sendEmail, not sendMail
Here's what it does, run in a infinite while loop:
0) Kill any old instances of the miners that may be running (Cleanup) and start the while loop from #1 below
1) Fetch your control url
2) If #1 fails, sleep and repeat the main while loop
3) Check if miner is running, if not, start it using the address from step #1 above
4) Check if the address currently in use is different from #1
5) if so, kill the currently running miner (when we loop next time, we start with new address)
6) If address has not changed, check if Temp is under limit, if Temp > Threshold, kill miner
7) Check Load, if load is under threshold, sleep 15 seconds
8) Check Load again, if load is under threshold, kill miner, (When loop repeats, it will start miner)
The recheck period is 10 seconds, and we recheck only under these conditions:
* If miner wasn't running and we just restarted it
* If the miner was killed because load was under specified threshhold
Aside from those two conditions, the loop repeats in 300 seconds (5 minutes)
The configuration settings are at the top of the script, here is a bit more detailed description of everything
CTRL_URL - this needs to point to something that returns a phoenix-friendly URL
MINER_ID - this is an idintification string to differentiate different screen sessions for people that have multiple GPU's in one system
ATICONFIG_ADAPTER_ID - this should contain an integer as you would specifiy for --adapter=X when running aticonfig for checking Temperature and Load
PHOENIX_OPTIONS - this is an entire string of commands that you pass to the phoenix miner
MIN_LOAD - if the load on the GPU drop below this limit, we'll kill the miner
MAX_TEMP - if the temp reported by aticonfig goes above this limit, we'll kill the miner
ENABLE_EMAIL - set this to 1 if you want to recieve spam/email alerts from your miner
# some basic email settings
ENABLE_LOG=1; # 1=enabled, anything else will disable it