HG stops responding

[jshan]

Hoping someone can help...

I'm running HG 493 on a rPi.  After running for about a week without any issues, HG stops responding.  The log is clean - no errors.  I can putty into the rPi and restart the homegenie service.  I could automatically restart HG every few days but I'd prefer to figure out the issue.  But I don't know how to start.  Can anyone offer suggestions as to how to proceed?  Thanks.

[petediscrete]

Might help if you tell us what you are trying to run on the RPI ie x10, Z Wave, sensors, cameras etc.

[jshan]

Thanks for the response.  I'm running an Aeon Z-wave USB dongle communicating to a z-wave siren and motion sensor.  I have a USB wifi dongle.  I have a USB to RS232 dongle with a CM17A X-10 controlling 4 lights (very seldom used) using an app as described here (http://www.homegenie.it/forum/index.php?topic=528.0).  I have 2 Foscam cameras feeding into HG via wifi.  I'm running CheckIP which runs every hour.  I run ShowTimeDate as described here (http://www.homegenie.it/forum/index.php?topic=1024.0).  I have a DHT-22 hanging off a GPIO reporting the temp/humidity and I have a magnetic sensor also hanging off a GPIO.  I use WeatherUnderground and the alarm system regularly.  I have the rPi e-mailing me from CheckIP and if an alarm occurs.  I run Android Push Notifications as described here (http://www.homegenie.it/forum/index.php?topic=541.0) if the alarm triggers.  I have the scheduler app running but not triggering anything.  I think that's generally about it that.  I could start turning stuff off to see if the issue goes away but because it takes about a week, I could be doing this for a long time, so I'd prefer if there was a way to determine why HG stops - is there a way to query Mono after the fact to see if it ran out of memory or something else?

[Gene]

Hi jshan,

a possibly related fix has just been done in latest source code. If you don't want to build from source, I think I will release the new hg package before the end of this week.

https://github.com/genielabs/HomeGenie

The problem was happening after a certain number of HTTP requests made by clients not closing the connections.
This can be verified by issuing this command:

Code: [Select]

sudo netstat -tupn --all | grep mono | grep tcp

if you see lot of connections in the CLOSE_WAIT state, then this is the bug.

Cheers,
g.

[jshan]

Thank you, Gene!  I had to restart the HG service yesterday again.  Typing in the netstat command you suggested, there are 5 lines reported, 3 "Listen", 2 "Close_Wait".  I look forward to the fix.  Thanks again!