Restarts on just about any scene or Reactor

Great points here about running it more naturally.

I removed the ecobee activity, which makes the list of activities only:

  • actual devices
  • reactor variable setting
  • delay / timer
  • actual devices

And it still triggered a restart. Next I’ll put the ecobee mode back in take the reactor variable out.

1 Like

So when I took the reactor variable setting out of the activities, the group executed (device actions, ecobee mode setting, delay timer, device actions) without a restart.

I went to add the variable setting back in and see that it’s captured and displayed differently than previous. @rigpapa Was there a semi-recent change in how expression/variable setting was structured?

I removed and then re-entered it as both an action and condition, reran the group activity by triggering it false and then true again, and everything ran correctly with no restart.

Related note: I’ve got a reactor variable set action in a scene, and that doesn’t seem to have any difference in structure but wondering if that’s failing and causing that scene to trigger restart.

I don’t know what this means. Can you explain? Examples? Before/after?

No, unless you’ve manually updated your Reactor installation, there have been no new releases.

Ah ahahahaha… wait a minute… you had a “Device Action” using the “Set Variable” UPnP action before. Did you choose a “Set Variable” activity action the second time around… that is, in the leftmost field on the line, does it say “Device Action” now or “Set Variable”? If the latter, that’s a different action from what you had before! It will do the same thing, but it’s also MUCH more efficient.

exactly! It is much more simple in layout. Here is the comparison, but they’re in different sensors, so disregard all the different values…

Old/Original method:

New/Simpler method:

I will say though, that the top one, which is the original method, is still working fine.

The new version is really more efficient, because it’s a direct action on the variable. The “Device Action” works, but it sends the command the Vera/Luup, which then sends the command back to Reactor. There are some other small differences, but that’s the meat of it. But the fact that it now doesn’t get submitted to the Luup engine means one less opportunity for job control to go wrong.

2 Likes

You guys jinxed me! Tonight, for the first time in a long while, my Vera restarted right after my customary “Lights Out” and “Nightlight” routines ran. I’ll have to investigate tomorrow.

The vera bug must have migrated - this morning was the first time in weeks the automation didn’t trigger the restart. I will add that it did last night, which is an instance where I have the reactor variable being set from within a scene. I’ll try to figure out a way to get this action into the reactor, but I have it this way now in case I need to run the scene manually from Imperihome or something (without reactor) then the variable is correct for the morning scene to work properly.

Is there a method to ‘search’ for all instances of the variable like this? I could go into the logic report for each sensor and do a ctrl+f… but what about scenes, etc.

Are you aware of anything I could do to perform a ‘global’ search of sorts?

HEL-LO! Upon closer inspection the following morning, I happened upon a minor train wreck inside one of my Reactor sensors, which I’ll try to tease apart below:

  1. A Stop Sign warning with lots of red ink, signifying “missing” references (see screenshot, below);
  2. Reactor would not allow me to “correct” those fields, instead only offering “Undefined is true” or “Undefined is false” as options;
  3. Over in the referred RS (Hallway), those desired Activity Groups still exist (and have no red ink)!
    Even weirder, to me, is the fact that Reactor has done the expected thing every time I press the Nodon remote button (e.g. ‘Minus’ pressed once) in question, namely it runs the activity group (grpo5qyxyk.true) “Decrement A/C Temp” just fine.

    But I do believe that button-press coincides almost exactly with the Luup restart I mentioned previously, as its time stamp (I have Reactor set to Notify me via SMTP email every time Vera restarts) aligns closely with the time of night I pressed it.

Not sure how best to proceed (i.e. how to get the Stop Sign warning to go away and be able to select the desired [Run Group] activities)? For the record, I haven’t edited either of these Reactor sensors nor updated Reactor versions in several weeks.

Also for the record, this sequence involves both setting (and using) Expression variables as well as interacting with the ecobee plugin!

I’ve seen similar where the device is missing - for one reason or another. That I’m sure is on Vera’s end… The trick to getting the red warning dialogue corrected is to delete the row and create again. It will re-retrieve the device type, choice of actions, etc.

1 Like

Ah, but you see, that’s part of the overall problem. Even removing and attempting to re-add the [Run Group] line (as shown in this screenshot) results in this hangup:


That is, I can only choose from two equally ‘wrong’ (undefined) options. Meanwhile, the target RS (#183) indeed has many Activity groups intact, yet I cannot select them here.

I’ll try removing all … saving … adding back all. Hold on…
ANSWER: NOPE!

Yes - good idea - and you can also try restoring just that one sensor from backup.

1 Like

Since delete-save-retry didn’t work, I’m doing a hard refresh and then will make 2nd attempt to re-add those [RUN GROUP] lines…

…sadly, NO GO. Further experimentation shows that I cannot even add a [RUN GROUP] referencing that (Hallway) RS to any other RS. So I’m led to believe the problem essentially lies with (Hallway) somehow.

I’ll try rebuilding (Hallway) in parts, or doing a RESTORE for it alone (this is what I tried first), and see if things improve after another hard refresh…
…NOPE!
…and a Net > Reboot…?
…NOPE!

1 Like

I’m summoning @rigpapa at this juncture, as I’m now finding that it’s impossible to add a [RUN GROUP] line to any RS, referencing any other RS. Same ‘undefined is true/false’ choices always pop up.

Is somethin’ borked?

P.S. Hadn’t intended to hijack this thread! Thought I was onto something helpful at first, lol.

1 Like

Shouldn’t matter but have you updated the device database?

Can you add another Reactor and reference that correctly? I wonder if there’s a storage issue.

Still, that’s a couple of wrong answers to keep you busy until Patrick arrives with the correct one :wink:

C

2 Likes

Intriguing suggestions (thanks), but with odd results, in that clicking “UPDATE DEVICE DATABASE” under “Tools” caused the “Please wait, downloading update” message to display and remain for several minutes (see screenshot below).


It still hasn’t gone away. Thus can’t tell if all this is a “Me” problem or a “Reactor” problem, but I’m halfway tempted to reinstall the latest Stable branch of Reactor from Github if things don’t settle.

Hmm, not at all sure about that. Could well be a ‘them’ issue I guess.

I’m pretty sure that re-installing it won’t help. Have you tried creating a new sensor and seeing if you can
a) Reference the old sensor that you get the ‘undefined’ message for
b) Reference the new sensor in the one with all the red ink?

Just thinking aloud really

C

1 Like

As a big fan of the scientific method, I followed your wise suggestion of adding yet another RS (my 14th), then did the “dance”, but sadly found it, too, unable to reference any Group Activities from any other RS – or its own, frankly (the drop-down is completely blank when I try that) – so overall [RUN GROUP ACTIVITY] remains borked for my setup. :frowning:

1 Like

Very odd

Space OK? (See how I’m keeping you warm for the expert) :wink:

C

1 Like