Indigo Domotics

Hi, I have the problem that the indigo hangs quite often (spinning ball - no z-wave or inputs at the client interface are functional). Sometimes it comes back to normal after some minutes, sometimes only force quit and restart of the client helps. I read in the forum that it might to have to do with plugins.

I would like to check the logs first for certain errors. What is a clever way to do it? Since the logs are several text files its quite tedious to search them ...especially not sure for what I should search. Is there a clever way to do this? Eg a text editor that highlight certain type of messages or makes a summary how often messages occur etc? Or to find out how long the client is unresponsive.

Is there any incompatibility known on a mac mini mid 2012 running 10.13.6 that could explain this kind of problem?

One thing to add, I also experience quite often that I get the error for several devices that "does not support off command" when trying to switch off a light e.g. This error can be solved by restarting z-wave.

Appreciate any ideas.

Is it just the Indigo Client hanging, or do you also have to restart the Indigo Server to get it working again?

The client can hang or get into a bad state if you have plugins that are updating device states or variable values too quickly for the macOS to handle the UI updates. A good approach in that case is to disable some of the plugins you are using to try to narrow down where the problem is coming from. If it from Device state updates and you have all of your devices in the main root level (not in any folders) then it can help to divide them out among several folders so that not all of the devices are shown/visible in the Main Window at once.

superholz wrote:

Is there a clever way to do this?

If you aren't seeing errors in the Event Log that coincide with the problem times, then the Event Log may not help. You can copy/paste chunks of it that coincide with the problem to a follow-up post if you'd like us to see if anything looks suspicious. The key though is that to be useful we need the lines that coincide with the problem.

superholz wrote:

Is there any incompatibility known on a mac mini mid 2012 running 10.13.6 that could explain this kind of problem?

Nope.

The next time it i hanging, open the built-in macOS app Activity Monitor and see any of the Indigo processes are using a lot of memory or CPU.

Hi Matt,

I seem to have the exact same promlem. and it's not new. The Indigo Client Window is not movable and nothing else seems to do anything. After about 10-15 seconds it reacts again and freezes a short time after that again.
The activity monitor shows noting special.
It seems to me the main loop auf the client hangs.
I am still using 7.3 btw.
What can I do to debug this?

Wilhelm

Additional observation:
Memory consumption of the kernel task has risen from 1 GB after reboot to 5,5 GB during the last week.

I have disabled all plugins, but the freezes are still there.

Did you quit and restart the Indigo client and Server and disabling the plugins? Once it gets into a bad state, I'm not sure just disabling the plugins will recover it (so restart everything w/ the plugins disabled and see if the problem persists).

I disabled most of the plugin and restarted the server.
It came back just the same.
What I don’t have is the beachball.
Maybe because the server has 16GB of memory.

The server itself is responsive, just not the Indigo frontend...
Is there any debug logs that I could enable?

Debug logs are unlikely to help in this case. In the Activity Monitor app what are the specifics memory and CPU stats shown for the Indigo client (when it is bogged down)? Also helpful to see the Indigo Server stats at the same time (even if the Indigo Server if functioning okay).

Normally I have the client app running on the server.
I will try using the client remotely and report back.
Thanks

I have the client now running on a different machine.
It still happens, that the client window freezes on a regular basis. To test this, I just try to drag the window arround.
As soon as the Indigo 7.3 Client Process take more than 10% of the CPU the window cannot be moved.

On thh server side eveything seems to run smoothly.
So the servers now takes about 5-10 % CPU and the Client App takes between 0.4%, and 35 % (which I find odd btw.)
I will keep the client running and watch if the kernel_task memory now increases on the client.

It seems to me, the problem has to do with the screen refresh of the client window.

If I minimize the window, CPU never goes over 9,5% on that machine.
If I reopen the window and the client displays one of the list, the cpu load immediately jumps up to 35% and the window frezzes for some time.

Does it occur with all of Indigo's windows or just Main Window? If just the Main Window, what happens if you select different folders/items (so Control Pages view versus Devices, etc.) and then try dragging it around?

Hi Matt,

yesterday I was pretty sure it was just the main window.
Trying to verify this today leaves me puzzled, because I cannot reproduce the issue...
The only thing I found out today:
If I select all devices in the list at once, cpu jumps up to 100%, activity monitor says "not responding" and the client app hangs until I force a termination via activity monitor
But that's not really what the problem was in the first place.

As soon as I have better information I will report back.
Thanks
Wilhelm

It sounds like you are running into the problem / race condition where you have enough devices updating at a frequency that the macOS cannot keep up with the UI updates. Normally this is because of a script or plugin that is updating a lot of devices frequently.

I had that problem w pibeacon,, # of sensors > 100 that were updating @ 10+ sensors / second.. that caused problems not just for indigo - the display was hanging... had to reboot the server frequently

THAT plugin was running at 80% of one cpu., total cpu though was < 15%.

Suggestion: try "utilities" plugin, that gives you a nice handle / info of
(A) trigger on too much cpu per plugin in one measurement period
(B) can give longterm reporting on cpu usage

I fixed the plugin (to be posted later today), its now running at < 5%

Karl

I just updated my plugins and now the server is hanging and cpu bound. How can I disable the plugins if I can't get the server established?