OMNI Panel Offline Recovery

[ericg]

For various reasons, both accidental and deliberate, I have often experienced temporary interruptions in the Ethernet link between my HomeTroller and my OmniPro II panel. Whenever this occurs, the status of device OMNI System Connection changes to status PANEL OFFLINE (Duh!).

The only effective recovery I have found is to disable manually the OMNI plugin, then re-enable it. On my system, I'm back in business about 12 minutes later. Undesirable aspects of this procedure are:

1. I don't always discover the offline status until long after the occurrence.
2. When I do recognize the problem, I'm not always available to restart the plugin.
3. Sometimes 12 minutes can be a long time.

I have considered possible ways to improve the situation:

1. My first choice, if feasible, would be a mod to the OMNI plugin that monitors the panel status continuously. Whenever the panel went offline, it would automatically (periodically?) attempt to reestablish panel online status. Presumably, an automatic HS/OMNI device resync would be part of the recovery.
2. If this isn't feasible, I'd like to create an event that would trigger on OMNI PANEL OFFLINE status. The event trigger would cause the the OMNI plugin to be shut down and then restarted.

I do have EasyTrigger, but unfortunately, I don't know how to tell an event to shut down and then restart a plugin.

1. Can this be done at all?
2. If so, can it be done without scripting?
3. If it can be done, but only with a script, can anyone help me out? (I have not yet ventured into the world of scripts.)

Thanks.

Current Date/Time: 11/8/2018 1:29:01 PM
HomeSeer Version: HS3 Pro Edition 3.0.0.445
Linux version: Linux hometrollerSEL 3.16.0-031600-generic #201408031935 SMP Sun Aug 3 23:56:17 UTC 2014 i686 i686 i686 GNU/Linux System Uptime: 1 Day 19 Hours 54 Minutes 1 Second
IP Address: 192.168.1.105
Number of Devices: 177
Number of Events: 36
Available Threads: 199
HSTouch Enabled: True
Event Threads: 0
Event Trigger Eval Queue: 0
Event Trigger Priority Eval Queue: 0
Device Exec Queue: 0
HSTouch Event Queue: 0
Email Send Queue: 0
Anti Virus Installed:

Enabled Plug-Ins
1.0.0.147: Arduino Plugin
3.0.19.0: BLLock
1.3.4.3: Device History
3.0.0.51: EasyTrigger
3.0.2.18: OMNI
3.0.1.228: Z-Wave

[rmasonjr]

Yeah, 12 mins is pretty slow if you have a larger installation.
The ethernet port on the OMNI is an older technology that will exhibit connection issues if plugged into newer switches/hubs. If you can, try an older 10/100 hub between your OMNI and your switch. I run one on my installation and dont have any dropouts. Or, if you have a managed switch, configure it to 10Mbps/half-duplex.

The OMNI SDK has a built-in heartbeat that will poll the panel every 'n' seconds with an ACK to keep the connection alive. Along with that, the plugin has its own 15 sec heartbeat to keep alive. If you are having connection issues, I'd like you to try the hardware approach first.

Building code to reconnect might cause more problems. If you are having a 12m startup time, that would be multiplied if the plugin disconnects/reconnects all the time.
Lastly, post your latest logs and let's make sure nothing else is going on that needs addressing.

[ericg]

rmasonjr, I probably misled you somewhat. Under normal circumstances, the system runs for long periods without hiccups. I have read other postings about the OPII's problematic Ethernet interface. Although I am using a no-name 8-port 10/100/1000 switch in my security closet, it has been basically reliable. These are the known causes of my link disruptions:

1. To get from my HT to my OPII, the Ethernet signal has to traverse 2 unmanaged switches, 4 Cat-5 cables, and 7 Ethernet cable connectors. The tiniest glitch will drop my OMNI link. Although I have no data on this specific link, I have frequently observed somewhat flaky connections on my various Ethernet LAN connection points. I often resolve (if only temporarily) problems by reseating a connector plug, or even by moving it to another port.
2. I recently lost my link because my cat stepped on the on/off switch of a power strip that supplied an Ethernet switch.
3. I somehow managed to break physically the wall wart that supplies 12 volts to my security closet Ethernet switch. I had to break the Ethernet link twice -- once to substitute a backup Ethernet switch, and again to restore the original one after I obtained a replacement wall wart.
4. I live in a rural area that is subject to frequent power failures, any of which will also disrupt the link. I am in the process of adding UPS units to maintain uninterrupted critical power during short term mains failure, but installation and maintenance of UPS implies still more link interruptions.

To summarize, if nobody touches anything, and if Pacific Power keeps the juice flowing, the system appears reliable. But as an old friend used to say, "If frogs had wings, they wouldn't bump their a** every time they jumped."

I believe I understand your concern about any automatic recovery that might itself be sabotaged by additional glitches during the recovery attempt. An easy way to mitigate this potentiality is to inhibit the attempt until the link has proven consistently good for at least the time required to reinitialize.

Separately, do you know of any event driven plugin disable/enable control option?

Finally, thank you for a really fantastic plugin. Your work has made HS and OMNI working together a much more powerful system than I could create with either (or even both!) of them working separately.