Can a connection lost with myhs cause a memory leak?

[Guest]

Apr-04 00:09:20

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 00:09:20

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 00:09:09

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

[Guest]

Apr-04 01:13:15

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 01:13:15

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 01:13:04

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Apr-04 01:09:45

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 01:09:45

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 01:09:34

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

[Guest]

Apr-04 01:24:36

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 01:24:36

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 01:24:04

Warning

MyHS could not connect to server: Could not resolve host 'connected2.homeseer.com'

Apr-04 01:23:43

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

[Guest]

rjh ?

[Guest]

Te first one was not intentional, but the others one were.

[Guest]

Continuation from that post

I can say with 100% certainty that if you install HS fresh and just run it, it does not leak. We have over dozen systems running here and we don't see any leaks (Windows and Linux). So what this means is that there is something on your system causing the leak and I cannot fix it until I can re-create it. So I need your help to do the following:

* Disable all plugins, does it still leak? If not, its a plugin, enable each one until the leak comes back
* If it still leaks with all plugins disabled, disable all your events (you can do this in HS Setup). If the leak stops, enable each event until the leak comes back. Note if an event runs a script, it could be a script causing the leak.
* if it still leaks when plugins are disabled and events are disabled, then its probably an operating system issue and its worth trying your config on a different system. A corrupt OS can cause a leak.

I understand that the above process is a real pain, but its the only way to figure out where the leak is.

I have disable event processing in he config for now, and disabled all plugin.

will start with that first. So i will know what side to chek ... ( hs3/os/mono) or the other way ( events/scripts/plugins). ... ( or maybe both(hope not the case))
Depending on what side it goes, it could take a while.


So this is an almost a virgin hs3 running . The data is sent via mqtt ( 5 secs interval) to nodered for graphing ; will let this go for 1-2 days. :

I will upload/refresh the same picture your are seeing here.. from time to time. rjh
Then, in a next post I will disable myhs, json, and ASCII, if this test is not ''stable''(they are still enable).

[Guest]

Seems stable to 135-140 mb ?

[Guest]

Apr-04 21:07:15

Info

System connected to MyHS Service successfully with license ID xxxx.

Apr-04 21:07:15

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 21:07:04

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Apr-04 21:01:47

Info

System connected to MyHS Service successfully with license ID xxxx.

Apr-04 21:01:47

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 21:01:36

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Apr-04 20:26:09

Info

System connected to MyHS Service successfully with license ID xxxx.

Apr-04 20:26:09

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 20:25:58

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Will do couple of test by removing ethernet wire from the server.


edit : Did 3 tests, and they correlate to the graph... will let the system goes.... to see if it goes back to near 140 mb. And will remove ethernet cable again in 1 hour.

[Guest]

Apr-04 22:24:15

Info

System connected to MyHS Service successfully with license ID xxxxxx.

Apr-04 22:24:15

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 22:24:04

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Apr-04 22:17:30

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 22:17:30

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 22:17:19

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Apr-04 22:15:00

Log Info

The log database is currently 45.23MB in size.

Apr-04 22:11:24

Info

System connected to MyHS Service successfully with license ID xxxxx.

Apr-04 22:11:24

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-04 22:11:12

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

Going to sleep. will let this run for the night

[Guest]

The frist disconnections with myhs seem to have the greater impact. +- 20 mb

The seconds seem to have less impact +- 10 mb


edit: seems to go from 165-170mb now.

I think the act of retriving or searching for my hs is a small leak.

. tomorow when i come back home ~2-4 pm. I’ll chek if all is stable. Then try another connection drop. Then I will try to remove plugin by plugin as you said etc.

right now i’m on my phone . But from ssh i see that it has been at 176mb since the last 10 mins. Before that i do not know.

[Guest]

here I have reactivated all plugin an the events engine. Will let this go for a while.


I think we can infer than myhs connection drop can increase ram by 40 mb. (175-135 =40). Maybe more if the delay between the test were larger?

[Guest]

I think I found the leak! rjh macromark

When I saw the memory leak, i was changing from z wave v.252 to v.260,. then reverted back... then went again to v260. Was tring to resolve the issue with the z-wave stick.


In the graph below, as you can see, with all the plugin loaded, and the event engine on, it is pretty stable at 220 mb.

But when i did the process of changing from zwave v.260, to v252, back and forth... you can look what is happening on the graph.

So i guess... overtime, someone would change is plugin version ( z wave or other plugin) and overtime someone my be out of memory on day,..

and with the myhs connection drops , it is another 20-40 mb that can add up.

[Guest]

The last spike you see , here at 17:31, is when doing the same with Easytrigger plugin.

I will of course restart my sever now.

[Guest]

a last proof of myhs leak ( this one was not on purpose)

Apr-05 19:15:00

Log Info

The log database is currently 46.34MB in size.

Apr-05 19:12:32

Info

System connected to MyHS Service successfully with license ID xxxx.

Apr-05 19:12:32

Info

System connected to MyHS Service, waiting for acknowledge...

Apr-05 19:12:21

Warning

Have not heard from MyHS in over 3 minutes, disconnecting for re-connection.

[terrydeadwyler]

How was sending data on memory usage setup for this data?