bpwwer

Fixed the name error in 2.0.14 @Lore I don't see any errors in the log but I also don't see that it ever called the initialization function so I don't really understand what's happening.

Looks like I forgot to push the changes from my local copy. I just did that so it should update to 2.0.13 now. Set the log level to debug and then download the log an either attach or PM it to me.

The problems looks like it's happening in the UDI node.js interface library. I'm guessing they never tested it with the upgrade to verify that it works. I'm no longer maintaining that library so you'll probably have to contact UDI.

version 2.0.13 has the additional info in the log message

That's where the log stops? After "Create Roomba Object" it calls into the external roomba library and tries to initialize the robot. One of the very first things it does it output a log message, which isn't in your log. Can you check the PG3x log to see if there are any errors there? The only thing I can think of is that it's calling in with bad parameters. I'll add a log message to log what parameters it's sending.

Sorry, my mistake, fixed and submitted as version 2.0.12

What I see in the log is that the Emporia server is returning 500 - internal server error for all the requests between 12:15am and 7:25am. During that time nothing will get updated since the server isn't responding with anything but the error. After that, the server starts responding with data and continues to do so through the end of log. It doesn't look like the plug-in was restarted at 7:25. A 500 error is a pretty generic error. It simply means that the server encountered an error while trying to handle the request. Since the server is able to handle the requests from the plug-in normally, this tends to rule out the plug-in sending an invalid request. It's possible that the Emporia server is overloaded and doesn't have the resources to handle the request, but I believe there are other errors it could return to indicate that. Most likely is that the server handling the request simply can't connect or access the data from their internal database, which generates an error and that's passed back to the plug-in. Most of the errors showing up the log are coming from the libraries used by the plug-in to make the request. The plug-in is throwing one error too, and I could trap that, but it won't be able to handle all the other errors that get logged when the server doesn't respond properly. From this log, I don't see anything that would indicate that the plug-in would be causing any resource issues with PG3x or IoX. In fact, when these errors are occurring, it's putting less load on PG3x and IoX because it's not sending any status updates to them.

WeatherBit is a bit of a pain right now. More than one person (including me) have had issues where our API key is fine, we're not exceeding the rate limit but it will still return API key not valid to queries. There's nothing I can do from the plug-in side to fix that, it's all on WeatherBit's side and they haven't been responsive to support requests about this. My best guess is that they really don't want to support the 'free' plan anymore, but who knows.

PG3x has a log file that you can access from the main PG3x menu (just called 'Log'). This log will show any errors when starting a plug-in that occur before the plug-in actually starts.

Submitted version 2.0.11 to the store. This has an updated getPassword() function that should be compatible with the latest OpenSSL versions.

Yeah, all my fault. Made what I thought was a simple change and didn't test it. Fixed in 2.0.10 which I just submitted to the store.

Ok, with multiple people seeing the same thing, it probably means that I screwed something up. I'll take a look and report back.

The plug-in log viewer can't show historical log entries. When you click on the log tab, it connect to the log file and then will start showing new entries as they are output to the log. If nothing is showing up, it could be because no new entries are being logged. In some cases, when the log tab isn't working, I've had to refresh/reload the log tab to get it working again.

Very rarely, or if PG3x crashes it can fail to kill running plug-ins. From your description it sounds like that's what has happened here. There is a copy of the Roomba plug-in running but PG3x thinks it has stopped that copy. Only one copy of a plug-in for a slot can be running so when you try to start a second copy, it can't and nothing happens (other than the message that it is already running). Depending on your comfort level with FreeBSD command line, you can either ssh in and kill the running copy or simply reboot the eisy/Polisy. Once that first copy is stopped/killed, you should be able to start it again from PG3x.

I don't know then, if it used to work, nothing has changed on the MH plug-in. Maybe try viewing what's happening when you press the keypad button using the diagnostics -> event view on admin console. That should show the incoming message from the keypad button and what the IOX sends out in response. If your system is busy doing a lot of stuff, it may be hard to pick those out of the view as it will also be showing all the other on-going activity.

bla I think it has something to do with the way IOX creates and handles scenes. An Insteon only scene programs the responders to respond to an Insteon broadcast message sent by the controller. In a mixed scene (Insteon + plug-in based device). That won't work because the plug-in based device (MH) doesn't use the Insteon protocol and can't respond to the broadcast message. IOX should know that the scene is a mixed device scene, but I don't think you can program a keypadlinc button to send anything but broadcast message. The keypadlinc will only send direct on/off commands for the main buttons. So that probably explains the behavior. When IOX tries to control the scene, it sends the proper command (may a broadcast) to control the Insteon devices and sends an ON command to control the MH device. So it all works as expected. But the keypadlinc only sends to broadcast command which controls the Insteon device but not MH device. The only workaround I can think of would be to supplement the scene with a program. If keypadlink button X is on, send on to MH device. This would probably be something you should submit to UDI as a ticket. They can then investigate if there's some way for IOX to handle this internally.

I'm seeing failures every few days where it seems like the device just stops responding. I'm connecting to the MQTT broker on the device itself, not remotely via the server. I haven't been able to investigate it yet, but it may just be badly designed hardware (well, probably the device firmware)

It appears that you don't have the updated roomba.py file. The failure is showing on line 42, which in the latest version is a blank line so very likely not a failure. The current version of roomba.py is 2.0.2, version 2.0.0i would have that failure on line 42

Never mind, that took about a minute to do so I just submitted version 2.0.9 to the store

Looks like that's another place that needs the workaround to enable unsafe renegotiation. When I tested, it didn't call that function, probably because the vacuum was already know to the plug-in. I'm gone for a few days so it'll probably be sometime next week that I can fix it.

I managed to find a couple of things and it seems to be working for me now. First, it looks like the newer OpenSSL liibs now being used on eisy/Polisy block the security methods Roomba uses and I wasn't able to get a connection to the Roomba. I found a workaround for that overrides the block and that allows it to connect and I was getting data from the Roomba. At some point the external library being used switch to an asyncio based library but plug-ins are not asyncio. This means you have to some modifications in the plug-in to properly call the asyncio based functions in the library. I had already make most of those changes. But there was one place I didn't have it right and it was hanging in the initialization code where it waits for the Roomba to send out the updates to get the full state of the Roomba. Once I fixed that, the node was created and I think it's working. However, there still could be places where I don't have proper code in there to access the asyncio functions. Version 2.0.8 should be in the store now.

Not really orphaned, just low priority and difficult to debug as I'm not the original author of the plug-in and it does depend on the external library for all the interaction with the Roomba. Since that external library was developed without any documentation or support from Roomba, it is also possible that Roomba has updated the firmware so that it no longer works. The external library hasn't been updated in about 4 years except for one minor change and I've added that one change and it didn't make any difference. I'll spend some more time and see if I can track down the issue.

PG2 didn't support any time of licensing/purchasing of node servers. That capability was one of the major features added by PG3. If you paid something for a PG2 node server, that would have been between you and the developer and any license would be incompatible with PG3's licensing scheme. However, I am not aware of any PG2 node servers that weren't free, PG2 had no way to restrict loading and running of node servers. While some PG2 node servers were ported to PG3 with only minor changes, it's not possible to run PG2 node servers on PG3. Even for those that were simply converted to run on PG3, you have to install the PG3 version. The initial migration process from PG2 to PG3, would look at the installed PG2 node server and attempt to pre-install the PG3 version, that only worked in very few cases.

This isn't a plug-in issue. The plug-in sends a map that links the numeric value to a string to IoX. IoX won't read this until it is restart (at least prior to version 6.0 it wouldn't). Until it reads this map file for the plug-in , it will only send out the numeric value as that's what the plug-in is sending to it. This mainly effects UD mobile. The admin console also reads these plug-in map files when it starts. This is way most instructions for installing plug-ins say to restart the admin console after installing the plug-in. How the new 6.0 release of IoX effects all of this, I'm not really sure.

I think you should have included this information along with a list of features that are not yet present in the eisy-ui in the release announcement. I'd also be interested in seeing a time line when each feature will be added (roughly, like maybe in each quarter of the year). About the only thing I see eisy-ui currently doing is replacing the PG3 main menu and showing nodes and programs. I'm not really convinced that the way eisy-ui shows nodes is actually an improvement over the way the admin console does. Also, I wouldn't classify this as beta software. Beta typically means feature complete although not all features may work correctly and there are certainly bugs. By your own admission, this is not feature complete. What's there tends to work pretty good. I like how the variables are displayed and modified. However, I was a bit shocked that I couldn't even look at programs. I am looking forward to seeing how this evolves, it does seem to be a good start.