Re: [mh] WARN: The PLM did not respond to the last command

[Jeff S. <ne...@si...>]

On 05/27/2014 04:53 PM, Lieven Hollevoet wrote:
> Hey Jeff,
>
> I don’t know if this applies to your system, but have you seen this:
> http://stackoverflow.com/questions/15343490/cant-read-data-from-serial-port
>
> Might be a combination of the kernel version and the driver?
>
> Kind regards,
>   Lieven.

Sorry for the long delay on this -- been on vacation and otherwise occupied.

So to answer your implied question, I believe it is possible my system 
does have this issue.  I am running SL6 which is 2.6.32 but it is 
unclear to me (maybe someone else can tell me how to figure it out) 
whether this patch was ever applied to the Red Hat kernels.

Either way I don't think this is necessarily everything that is going on 
with the PLM since the system logs about "FTDI USB Serial Device 
converter now disconnected" are far less frequent than the the "PLM did 
not respond errors in MH"

In terms of patterns for the PLM reopening, I continue to scratch my 
head.  I last power cycled the PLM/restarted MH on June 30.  There were 
51 re-opens within the next day.  Then I went on vacation and there were 
no more reopens for 12 days at which point there was one.  4 days later 
later there was another.  2 more days and there were two more.  This 
pattern of one or two resets every day or two continued for a few more days.

So far two things seem to be happening:

1. There were a ton of reopens when the house was occupied but they 
largely stopped when the house was vacant.  This is almost certainly 
because the motion tracking and related occupancy/presence controls for 
lights which were causing a lot of additional network traffic.

2. In a more/less steady state environment there seems to be an increase 
in reopens over time.

Then vacation was over and the house became occupied again.  Strangely 
there were no reopens the first day back.  This isn't what I would have 
expected.  The 2nd day back there were 14.  The 3rd day back there were 
12. But then there were none for the next three days

Not sure if any of this is helpful in solving the PLM not responding issue.

Would be helpful to turn up insteon debugging and then record the data 
in/out of the PLM around the "did not respond" being triggered events? 
I don't know how long MH waits to receive an ACK from the PLM but maybe 
the data is coming a bit late and MH is just impatient!

Thanks,

Jeff