New Error on .460

[Ltek]

bdickhaus I don't have any,never had, timer devices. I'm on .460 (gave been since it was released) but just past day or so seeing tons of these.

rjh are you working on a fix?

[Mr_Resistor]

I agree it's likely plug-in related. I have no timers set. Would be nice if the "limit" introduced in .435 was configurable either through the UI or through a line in settings.ini.

[bdickhaus]

The problem isn't new, it has always been there. It is that one or more plugins are not responding quickly enough to the event change call backs. It was causing out of memory errors and other performance issues. In version .435 there was a limit placed on the number of event change call back requests that have been queued (not responded to).

I no longer see these warnings. I fixed it by eliminating all my timer devices (not the timers, the devices). What was happening was the timer devices were getting updated once a second and several plugins that had registered for event change call backs were not responding fast enough.

It is a problem with one or more plugins, but I agree that it would be helpful to know which one(s).

[Mr_Resistor]

I opened a ticket for this as well. Support got back to me with an ask to enable some enhanced logging.

"

You can enable some advanced debug logging for events processing from Tools > Control Panel. From this page, you will want to enter the following: hs.DebugMode = 2

Once you enter this and click Execute Command, you will see a "&" appear before the command from above. You can now navigate to View > Logs and see what is happening there.You can disable the debug logs by entering hs.DebugMode = 0 "

The log did reveal that this was happening more often that I thought, but I wasn't able to find anything. I sent the logs in and they're being analyzed.. Please note, not my system, enabling this level of login started producing between 1500 and 1700 log entries PER SECOND.... so I only sent them 3 minutes worth of logs, (which was over 11mb.).

[simonmason]

I am getting this error and I am running the released version .435. According to the forum this was released on August 28th - about when I saw the error appear. It is really annoying - anything that impacts the stability of my system impacts my life! Opening a support ticket now.

[Mr_Resistor]

Still seeing this on .473.

[Mr_Resistor]

Thanks for the additional info. I think I have plenty of free resources, more than enough compute and memory, but I also have tons of plugins (72) and a half-ton of devices (4885).

[shill]

Whenever I'm working on my Device History plugin in debug mode (running it from my development environment as a remote plugin) and hit a breakpoint, this message shows up in the log after I've left it in "pause" for a while, so I've assumed that since I'm registered for event change callbacks I'm blocking the queue with messages that aren't getting delivered while I'm paused.

That could be extrapolated to a situation where a script or plugin that is registered for event change callbacks isn't processing them fast enough, or on a server with insufficient resources for the plugin to execute as fast as the changes are coming in, or in systems with very "chatty" devices that change frequently, this is more likely to happen.

[Mr_Resistor]

Hopefully this will get some attention in a future release.

[bdraper]

I still see these errors a couple of times an hour, even with the plex-related plug-ins disabled. I have no timers running. Would be good to see this "fake news" corrected in a future release. More info in the log would help troubleshoot, but a bigger queue would maybe resolve.

I see this error a lot as well. I agree more info in the log would be great and a resolution would be even better.

[Mr_Resistor]

I still see these errors a couple of times an hour, even with the plex-related plug-ins disabled. I have no timers running. Would be good to see this "fake news" corrected in a future release. More info in the log would help troubleshoot, but a bigger queue would maybe resolve.

[bdraper]

Originally Rich said he'd identify the sources, hopefully he'll get to it. I emailed the help desk months ago to ask as well.
As it stands, setting an arbitrary 500 limit without taking into account the machine it's running on, and not identifying the culprit(s), is just more fake news

Z

Fake news... LOL, good one.

Since mine seem to be centered around the Scheduler Timer... I took a look and only have 20 timers and only 8 are running at the moment. I hope we are able to identify the source in a future release.

[vasrc]

Originally Rich said he'd identify the sources, hopefully he'll get to it. I emailed the help desk months ago to ask as well.
As it stands, setting an arbitrary 500 limit without taking into account the machine it's running on, and not identifying the culprit(s), is just more fake news

Z

[bdraper]

Disabling both PlexMonitor and BLPlex worked. Almost error free. 2 occurrences (3-5 messages) in the last 30 minutes.

It makes sense though, Since they both generated over 500 devices.... the polling was overflowing the queue. I'll shoot a note to Bob (Blade) to see if there's anything that can be done. Perhaps he can have the device updates split. Some device types update in one cycle, and another set update in a second cycle.

The other option is see if HS can up the message queue as others have been seeing the same queue overflow issue. When I see them, they come 3 to 5 in a row, which seems like it's overflowing the 500 message queue by a wide margin. Not sure if a queue of 1,000, 2,000, or even 3,000k is even possible.

I have a fairly small number of plugins and do not have any of the Plex plugins, but the message is still very persistent on my system.

When I turn on show source thread in the HomeSeer log, then all appear to be centered around a System even for a Scheduled timer.

(System)_28-Jul-2018 10:14:55.7303_Scheduler Timer

[randy]

Randy, you were one of the folks who chimed in on the 435 thread as having see this error, but only when the system was busy (editing events). Did you ever find a resolution? Did it just die down?

I'm still seeing the warning. It comes in a blast of log entries a couple of times a week. It is usually a blast of 10-30 log entries at the same time. I've had 127 of these warnings over the last two weeks. I have never been able to correlate their occurrence with any specific event, nor have I seen any problems at the time the warnings are logged. It happens mostly when HS has just started.