Investigation of Level Sensor Anomalies

[hamishmb]

Good news.

Just realised why rivercontrolsystem.log is doing that - it's set up to wait for Wendy Butts Pi and the Gate Valve Pi to connect before it proceeds.

[TerryJC]

Picky!

[TerryJC]

Thinking about this, I wonder if we couldn't come up with a strategy whereby SumpPi registers each device as it connects and flags up an error in the log if a device in config.py fails to register.

This would prevent the scenario where a failure in Gazebo Pi (for example) prevents SumpPi logging anything else for the rest of the system.

[hamishmb]

Yes, I'm sure we can do that . I just need to figure out how, and fix an annoying issue with the Sockets code - it will currently hang forever waiting for a connection, which isn't what we want.

[Penri]

Thanks Terry,

Have a great weekend.

[TerryJC]

Sitrep2:

After my weekend away. both systems are still running at around 20%. That's just over eight days now with no further problems.

Hamish,

I'm not able to do it just yet, but should we consider installing your latest version of the Software Framework and see what happens?

[hamishmb]

The new version I've been running here seems fine too, at least as far as not crashing, so I would say go for it. I assume you're talking about trying it on your test network, but we could deploy it at WMT if you like. I haven't implemented all the changes, being somewhat distracted with the NAS box, but there are a fair number of improvements.

I wonder whether the problem is damp, because it's seemed better at WMT since it stopped raining constantly.

[TerryJC]

Hamish,

I had a bit of time to kill this morning, so I pulled the latest version of the Software Framework and copied it on to my pretend SumpPi and onto the real SButtsPi. There is something amiss because it records 0 mm whatever I do with the float. (This is viewing the output to STDOUT on the monitor screen.) It's not the hardware because my test software works, so this needs a bit of investigation. I don't have time to look into it now because I have a dental appointment and then I'm taking my mother to WMT.

I may have time later to see what is going on.

[hamishmb]

I probably messed something up, because I split the HallEffectProbe class to separate the management stuff. I'll have a look later on, hopefully a simple fix

[hamishmb]

Okay, I can't see any obvious reason for this happening, unfortunately. I'll do some more poking around tomorrow and hopefully I'll figure something out. If not, I may need to come around to yours again Terry so I can debug it in-situ - I don't have a probe.

I'm wondering if it'd be good if we could have some kind of fake probe, either in software or hardware, to help me detect and debug these kinds of issue.