HA periodically losing Envisalink TPI connection

My DSC 1632 panel with Envisalink 3 has been running and solid for over 4 years now. I recently installed HA on spare Windows 10 Pro PC running HA as a service, and able to configure and get Envisalink alarm panel and sensors up and running fine. I do not (at least aware) of any other local network clients connecting to Envisalink, I do use EyezOn service which I understand to be separate from local TPI client.

The problem I’m having is that from anywhere from 4-18 hours after a HA restart, HA looses connection and doesn’t reconnect automatically. I have to restart HA to re-establish the Envislink connection. The errors when the Envisalink TPI connection closes in the HA log are consistently:

2017-10-10 03:23:39 ERROR (MainThread) [pyenvisalink.envisalink_base_client] The server closed the connection. Reconnecting…
2017-10-10 03:24:05 ERROR (MainThread) [homeassistant.core] Error doing job: Task exception was never retrieved
Traceback (most recent call last): {traceback left out here…}
OSError: [WinError 121] The semaphore timeout period has expired

I’m really bummed this doesn’t stay connected, as it is pretty cool to link DSC Envisalink alarm and sensor events in HA to IFTTT.com and other HA actions.

Let me know if there is a fix to this or others have seen similar problems. I can provide more config information if needed. thanks!

1 Like

When I don’t close the HA / Envisalink connection properly, I need to manually reboot the envisalink then restart HA. I think its because if HA doesn’t logout properly on the envisalink , it doesn’t allow its port open for when HA tries to login on during restart.

To reboot Envisalink, go to the envisalink web page then network page.

That doesn’t seem to be my issue, as each time HA is restarted it connects fine to Envisalink… only a half or so day later connection drops and HA log states that ‘server closed the connection…’ and HA doesn’t reconnect automatically to Envisalink until I restart HA again. Based on your suggestion though I will reboot Envisalink just to try and see if it helps any.

I rebooted Envisalink, I logged in via browser and selected Envisalink ‘Network’ page from top right… I then clicked on the ‘Reboot Envisalink’ text link at bottom of page but nothing happened, tried to click this a few times to no avail. So then I decided while I was there to change my Envisalink to static IP address. I typed in available static address outside router DHCP range and checked the ‘Make Network Settings Static?’ box… and then Envisalink then stated it was rebooting and rebooted. I also changed the Envisalink IP address in my HA configuration file.

After Envisalink rebooted and I restarted HA with new Envisalink static IP… HA has not lost Envisalink connection since! 4+ days now, hurrah! But unfortunately I’m not sure whether fix was the Envisalink reboot or changing Envisalink to a static IP? (no other HA changes made during these 4 days) But anyway Envisalink isn’t closing client connection with HA anymore (at least for now).

I’ll post again if Envisalink starts to misbehave again, thank you for suggesting reboot.

Do we need to check the Enable TPI Session Alerts ? or just leave it unchecked ?

I understand the use of the TPI Session Alerts checkbox is these events get sent to EyezON if you are registered for the free EyezON service, and configure to monitor your Envisalink. The TPI Session events can be viewed in your EyezON website Partition Recent Events view, and if you use the newer EyezON mobile app also, it can forward these as app notifications to your mobile phone (app setup auto configures a iPhone notification contact in your EyezOn config). As then you’ll get a mobile app notification right when the TPI Session closes on HA, or when HA logs into TPI Session. Useful to be aware of lost/closed HA Envisalink sessions. Be sure to configure in your EyezOn contacts to send most of the event types to the iPhone notification contact (or email or SMS if you prefer).

I have the same issue with my Envisalink 4. did the steps above solved your issue? @gudelhoven