I had a little problem when one of my Insteon devices wasn’t responding. It turned out to be just the tip of the iceberg. Attempting fix the problem through the integration utilities seemed to make matters worse.
Fortunately I still had my ISY994i from which I had migrated. I connected my PLM (serial) to it and began to explore. The PLM All-Link Database (ALDB) was extensively corrupted, as well as the database in a number of devices. I experimented with changing individual, but it was a rather frustrating and fruitess.
Start from scratch. The PLM requires the ability to hear a beep or guess the right number of seconds to hold the set button (15?) I’m pretty much deaf. I have no idea if I was successful. The ISY requires the most recent version of firm because it’s factory reset, resets to the firmware that shipped with it from the factory. An (obviously) inadequate search yielded no download and I really didn’t want to have to do an upgrade where, if things went bad, there was no support. Factory resetting the Insteon devices is a major pain.
Clear HA of all vestiges of Insteon: Delete, in this order, all scenes, all devices, and lastly, the PLM (which deletes the Insteon integration. I shutdown HA to ensure that the PLM would be seen
The process:
- note the Insteon address of all insteon devices.
- write down all the scenes and their constituents.
- delete each scene, one at a time. When a scene is deleted, the members of the scene must be update. This means communication between the PLM and the device, which takes a long time, 2-5 minutes per scene.
- I checked a few devices using the diagnostic, “Show Device Links Table”. Since there are no scenes, there should only be 2, 6 or 9 entries (switch, 6 keypad, 8 keypad, respectively). Some of the devices had over 50 entries! Fortunately, subsequent steps erased these
- delete, then install each device. This is slow and tedious
5a .delete device
5b. install device with erase device while installing
5c. check the box contents to ensure right number of db entries.
6.Backup the ISY - recreate the scenes.
7a. create a scene
7b. control click on all of the devices going into the scene, the drag the selection then drop on the scene icon.
7c. set whether a device is going to be just a responder and both, responder and controller. In my case, only the scene “All” had one controller. In all the rest, all the members were controller.
7d. Click on OK and play a few games of solitaire because the writing to all the members takes a long time and you don’t want to send any more traffic over the Insteon network because you risk of corrupting the data but definitely the process will take longer. - Once all that was done, I checked to see that all the scenes worked as expected
8a. check to see if all the members of the scene off
8b. Turn one of the devices on and check to see if all the other devices in the scene went on.
8c. Turn another off and check to see if all the others turned off.
8d. continue for each member, alternating on and off.
I have 11 devices and 8 scenes and this took about 5 hours.
Move the PLM back to the HA computational device and boot up. Once booted, install the Insteon integration. All of the devices and scenes should be discovered, but it takes a long time … Watch an episode of Extraordinary Attorney Woo.
For me, my patience was rewarded and my system was working as it was designed! So delighted in this, I thought I check for missing/broken links which there should be none.
Alas, there were two. which recommended reloading the ALDB. I did. and then there was another two. These recommended deleting them. No F…in’ way! This system works!
I moved the PLM back to the ISY and found no problems, though it did appear that HA had changed some configuration data in a device (or more?)
So why all this prelude? My experience has been if I had written, “Two broken links after complete reset, why is that?” my query would have been ignored for lack of information. (by the way, the system is up to date as of this writing)
I would like the authors of the Insteon integration to know that there are some ghosts in the machine. Here I have written a way to recreate one of thoses ghosts. As I noted above, these may just be the tip of an iceberg.