-
Notifications
You must be signed in to change notification settings - Fork 3
MVP_Debugging
The MVP comes with a validation script which exercises the main components of the code. It can be run at any time to confirm that things are set up correctly and running properly. If you encounter problems (things stop working), running the validation script is always a good first step in finding the problem. In MVP 1.0 it can be tested via: > bash ~/OpenAg-MVP/setup/Validate.sh In MVP 3.0 it can be tested via: > bash ~/MVP/setup/Validate.sh If all is correct, no errors will be seen. If there is an error, you will see it displayed in RED. The script runs starting with low level issues and works its way up through complexity. Always fix the first error encountered and re-run the script. Fixing low level issues will often take care of things higher up, and low level issues are almost guaranteed to create other problems (ie. if a sensor is not working, you will not log data or produce a data chart). The validation script gives little output. If you have an error, run the referenced script of code individually from the command line. For scripts: > bash <path></path><command.sh></command.sh> For python code: > python <path></path><code.py></code.py> You should also be able to click (or double click) on a python file to open in in the IDE and run it from there. These will give you the error details, which should point you toward what needs to be fixed. Though the code is not robust (having a lot of error handling), because it is simple and has been running for a while, you should encounter few (if any) actual code problems. Most problems have to do with the setup and configuration, often the problems are with typing errors in cron (which is not in the validation script).
The first step is to make sure the main directories are in place, that you didn't copy things from Github to the wrong location. If you encounter an error here, re-copy the Github download to the correct location. See the README.md for correct deployment.
This part checks that CouchDB is set up correctly, running, and has the proper database and view file. If the first test fails, it means the CouchDB is not running. Either it was not installed properly, or you need to reboot your system to get it running. The second test looks for the database where sensor data is stored. If it is missing, run the script to build it (see the README.md) The third test checks for the view script, this is needed to generate the charts. If it is missing, run the script to build it (see the README.md)
There are a couple of variables stored to help track the state of the MVP. 1.0 used the Python 'shelf' library, while 3.0 stores them in CouchDB. If this step fails, you likely did not load the variables (for 1.0 you may have a Python version conflict).
These tests perform low level reads of the sensors. If you encounter problems here, it is likely a wiring problem.
This runs the code that logs sensor data to CouchDB. If you have gotten this far, the sensors are working, and CouchDB is up; you are unlikely to see any issues here. This step assures there is something in the database for the chart to report.
When this test runs, you will likely hear your relay clicking, lights flashing and the fan going off and on. Software issues are unlikely, though this may show up some wiring problems. The USB camera will also be checked, by trying to take a picture.
Since it is assumed that there is data in CouchDB, and you have a new picture, this step runs > render.py Which will move the latest picture to the UI web directory and build the charts. Again, if you got to this point, there is little that can go wrong. Open your brower and check things to see what you have. > http://localhost:8000 For more detailed testing of the charting, see here.
It is always good to check these for errors or problems. For problems with cron, check the log: > /var/log/cron.log Ignore "No MTA installed". This is saying that you do not have email set up to send cron output messages.
For Server logs: In MVP 1.0 the log is at: > ~/MVP_UI/server.log In MVP 3.0 the log is at: > ~/MVP/logs/server.log
2018/05/31 webbhm