Upgraded to 5.1, lost "mon-from" messages from FHT80b

[Al]

Hi, I upgraded to 5.1, and now I only receive a few messages from the
FHT devices. Messages include actuator, window open. I do not receive
any of the timer on and off times.

I tried replacing 11_FHT.pm from the old version (5.0) and the full
message set appeared again.

Has anyone else seen this?

[Rudolf Koenig]

> I tried replacing 11_FHT.pm from the old version (5.0) and the full
> message set appeared again.

I just checked the diffs between the two versions, and I cannot imagine how
your problem could be caused by the changes. I suppose a coincidence with
another problem.

[Boris Neubert]

Am 08.10.2011 23:21, schrieb Al:
> I tried replacing 11_FHT.pm from the old version (5.0) and the full
> message set appeared again.

I would assume that the restart of fhem somehow triggered some
initialization sequence that made the FHTs more verbose again.

Regards,
Boris

[Al]

Thanks for the replies.
I've now restored the correct 11_FHT.pm from v5.1, and the problem has
returned.

Could it be that the report1 255 command is being ignored, and only
the report2 255 command is being sent?

Here is a paste from one of the FHTs. None of the timer or day temp
and night temp values have been received.

ack 0 2011-10-09 08:57:18
actuator 0% 2011-10-09 08:57:16
battery ok 2011-10-09 08:57:17
end-xmit 0 2011-10-09 08:57:18
lowtemp ok 2011-10-09 08:57:17
measured-high 0 2011-10-09 08:57:17
measured-low 172 2011-10-09 08:57:16
measured-temp 17.2 (Celsius) 2011-10-09 08:57:17
warnings none 2011-10-09 08:57:17
window closed 2011-10-09 08:57:17
windowsensor ok 2011-10-09 08:57:17

Internal Value
CODE 3204
DEF
3204
FHZ1_MSGCNT 22
FHZ1_RAWMSG 810c04ce0909a00132047e006700
FHZ1_TIME 2011-10-09 08:57:18
LASTIODev FHZ1
MSGCNT 22
NAME Bathroom_0
NR 9
STATE measured-temp: 17.2 (Celsius)
TYPE FHT

here is a paste of fhem.cfg

#####
#Global
#attr global autoload_undefined_devices 1
attr global logfile c:/FHEM/log/__fhem-%Y-%m.log
attr global modpath c:/FHEM
attr global nofork 1
attr global port 7072
attr global verbose 3
define WEB FHEMWEB 8083 global
define WEBphone FHEMWEB 8084 global
attr WEBphone smallscreen
define WEBtablet FHEMWEB 8085 global
attr WEBtablet touchpad
define Logfile FileLog c:/FHEM/log/__fhem-%Y-%m.log fakelog
define FHZ1 FHZ COM9
attr FHZ1 fhtsoftbuffer 1
define autocreate autocreate
attr autocreate device_room %TYPE
attr autocreate filelog c:/FHEM/log/%NAME-%Y-%m.log
attr autocreate weblink
attr autocreate weblink_room Plots

#Log
define A_OfficeLog FileLog c:/FHEM/log/_A_Office-%Y-%m.log A_Office
define Bathroom_0Log FileLog c:/FHEM/log/_Bathroom_0-%Y-%m.log
Bathroom_0
define Bathroom_1Log FileLog c:/FHEM/log/_Bathroom_1-%Y-%m.log
Bathroom_1
define BedroomLog FileLog c:/FHEM/log/_Bedroom-%Y-%m.log Bedroom
define HallLog FileLog c:/FHEM/log/_Hall-%Y-%m.log Hall
define J_OfficeLog FileLog c:/FHEM/log/_J_Office-%Y-%m.log J_Office
define KitchenLog FileLog c:/FHEM/log/_Kitchen-%Y-%m.log Kitchen
define LoungeLog FileLog c:/FHEM/log/_Lounge-%Y-%m.log Lounge
define Lounge_2Log FileLog c:/FHEM/log/_Lounge_2-%Y-%m.log Lounge_2
define Guest_BedroomLog FileLog c:/FHEM/log/_Guest_Bedroom-%Y-%m.log
Guest_Bedroom

#Define
define A_Office FHT 1415
define Bathroom_0 FHT 3204
define Bathroom_1 FHT 494c
define Bedroom FHT 1414
define Guest_Bedroom FHT 3d56
define Hall FHT 522a
define J_Office FHT 2a5e
define Kitchen FHT 5362
define Lounge FHT 3758
define Lounge_2 FHT 4252

#Retry
attr A_Office retrycount 10
attr Bathroom_0 retrycount 10
attr Bathroom_1 retrycount 10
attr Bedroom retrycount 10
attr Guest_Bedroom retrycount 10
attr Hall retrycount 10
attr J_Office retrycount 10
attr Kitchen retrycount 10
attr Lounge retrycount 10
attr Lounge_2 retrycount 10

#Lazy
attr A_Office lazy 1
attr Bathroom_0 lazy 1
attr Bathroom_1 lazy 1
attr Bedroom lazy 1
attr Guest_Bedroom lazy 1
attr Hall lazy 1
attr J_Office lazy 1
attr Kitchen lazy 1
attr Lounge lazy 1
attr Lounge_2 lazy 1

#Report
define A_OfficeReport at +*06:00:00 set A_Office report1 255 report2
255
define Bathroom_0Report at +*06:00:00 set Bathroom_0 report1 255
report2 255
define Bathroom_1Report at +*06:00:00 set Bathroom_1 report1 255
report2 255
define BedroomReport at +*06:00:00 set Bedroom report1 255 report2 255
define Guest_BedroomReport at +*06:00:00 set Guest_Bedroom report1 255
report2 255
define HallReport at +*06:00:00 set Hall report1 255 report2 255
define J_OfficeReport at +*06:00:00 set J_Office report1 255 report2
255
define KitchenReport at +*06:00:00 set Kitchen report1 255 report2 255
define LoungeReport at +*06:00:00 set Lounge report1 255 report2 255
define Lounge_2Report at +*06:00:00 set Lounge_2 report1 255 report2
255

#ReportNow
set A_Office refreshvalues
set Bathroom_0 refreshvalues
set Bathroom_1 refreshvalues
set Bedroom refreshvalues
set Hall refreshvalues
set J_Office refreshvalues
set Kitchen refreshvalues
set Lounge refreshvalues
set Lounge_2 refreshvalues
set Guest_Bedroom refreshvalues

[Rudolf Koenig]

> Could it be that the report1 255 command is being ignored, and only
> the report2 255 command is being sent?

There are only cosmetic changes between 5.0 and 5.1 for the FHZ+FHT
combination. You could try to remove the fhtsoftbuffer attribute, and check if
it makes a difference. But scheduling refreshvalues for 10 FHTs at once is
asking, or rather begging for trouble.

[Al]

> You could try to remove the fhtsoftbuffer attribute, and check if
> it makes a difference.

Ah, that did make a difference. When the fhtsoftbuffer is set to 0, I
get the full message set... Am I doing anything wrong?

[Al]

Is it normal for the fhtsoftbuffer attribute to affect the messages in
this way?

Thanks,
Al

[Al]

I still have this problem. with fhtsoftbuffer set to 1, then the 5.1
version of 11_FHT.pm does not receive the full message set. The 5.0
version of 11_FHT.pm works correctly. With fhtsoftbuffer set to 0,
then both versions work correctly.

Any ideas on how to diagnose further ?

thanks,

Al

[Rudolf Koenig]

> Any ideas on how to diagnose further ?

You are calling for ideas, here they are :)
- Trace what fhem writes to the FHZ, and compare the two versions.
- Monitor the air traffic between FHZ and FHT with a CUL in X61 mode.

Why are you sticking to the FHTsoftbuffer? It is only needed if you are sending
a lot of commands, which is due to the shaky communication between the FHZ and
FHT a problem in itself. Not that the CUL<->FHT would be any better.

[Al]

Thanks Rudolf. I don't have a CUL. As this is receive traffic, I'll
need to trap how fhem interprets the data from the FHZ.

I sometimes send several commands to the FHTs. For example I have a
spreadsheet that presents the from & to times for each FHT for each
day in an easy to read and update format. This is then easily
condensed into a list of "set" commands. I can then "include" a file
containing this list of commands into fhem to upload them. Lazy Mode
prevents most of them actually being transmitted.

As an example, changing the From 1, From 2, To 1 and To 2 times for 10
FHTs for 7 days is 240 commands. Yes I am aware I can combine multiple
commands on one "set" line, and I do this as much as possible.

With fhtsoftbuffer enabled, these commands trickle down to the FHTs
over a period of time, and this seems to work fine on version 5.0.

I appear to be in the minority in having problems with dropped receive
messages with 5.1 and fhtsoftbuffer enabled.

Thanks
Al

[Zrrronggg!]

I have no idea what's the difference between 5.1 <-> 5.0 and what
exactly causes the problem.

But Im really really certain, that already the

define FHTxy at +*06:00:00 set FHTxy report1 255 report2
255

stuff you do for TEN FHTs at the same time is calling for serious
trouble.
This should block the FHT communication for hours!

report 1 255 is the command causing the most traffic of all and you
are sending that to 10 FHTs at the same time, PLUS report 2 255 to 10
FHTs. There is no way, that this will work without problems. (Even if
you might not be able to really see them as you don't have a CUL)

From my experience I'm surprised that this even ever showed good
results.
I would assume that you even may run into problems with the maxium of
1% airtime ("Duty Cycle"). Putting retry to 10 may even worsen the
problem. This is a very high retry count.

Note, that sending single commands to an FHT usually happens only
every 2 minutes. So a retry count of 10 may consume up to 20 Minutes;
making it quite likely that FHT communications of multiple FHTs
interrupt each other.

Your method with the spreadsheet is also somewhat daring and probably
only works due to lazy mode.

To make this more reliable you may want to consider tuning this a bit.

May I suggest that you
1. spread the reports. Do the first room at f.e. 02:00:00 and the next
at 02:30:00 and then the next at 03:00:00 etc.

2. consider not doing the report1 255 every day but only once or twice
a week. Best spread the FHTs over the weekdays. I mean: do the daily
programs really change that often?
I use:

define fht_reportwz_o1 at *06:20:00 {if ($wday == 3) { fhem("set
hzg_wz_o report1 255") } }

3. consider doing report2 255 only every other day for groups of 5
FHTs or better spread them over the weekdays. Note, that the FHTs
usually send most of the report2 values anyhow every 15 Minutes. I
figured that I can even omit the whole report2 thing by using a
refresh-watchdog:

define wd_FHTwz_o watchdog hzg_wz_o:measured-temp.* 01:00 SAME set
hzg_wz_o time

This sets the time if the FHT does not send temp infos anymore, and
setting the time causes the FHT to send most of the report2 data
again.

4. lower retry count... a lot.

Let me explore retry count a bit further, if I may.

Why do you need that at all? Because some FHTs seem to not get the
commands, right?
This may have 2 reasons:
a) the radio situation is not optimal, the coverage is at its
borderline to some FHTs.
b) there is so much traffic in the air that packages cannot be
delivered undisturbed.

I do not know if a) is true with your installation or not. IF thats
the cause you should usually be abel to isolate the 2-3 FHTs which are
not covered well and then set the retry count for these up to...
say ... 4. Put the rest to 2.

Since ALL of your FHTs have a 10, I assume that's not your problem.
Then it is b).
If you set the retry count from (default) 3 to 10 you potentially more
than tripple the radio traffic. Having MORE of what causes the problem
in the first place might be not such a good idea. It does not only
worsen the traffic-problem, but also most likely leads to problems
with the 1% "Duty cycle". If that is reached, then NO communication
will take place at all for quite some time.
Using softbuffer guarantees that once the duty cycle waiting time is
over, so many commands are in the queue that the next duty cycle
problem is only seconds away. I do not know how the duty cycle is
implemented in the FHZ1X00 (which you probably use) but if it is done
similar to the CUL, then ONE "report1 255" will be already sufficient
to get problems again.

So basically I guess that the real reason of your problem lies in the
huge traffic you might have. Your large retry count makes me think,
that you already encounter the problems that may be caused by this.

So whatever the difference between the 2 versions is: In your current
setup, only minimal changes in the timing or behavior of the
softbuffer can make a huge difference, as I assume that your setup
runs at the very verge of not working.

[Al]

Zrrronggg,

Thank you very much for such a comprehensive explanation.

I like the watchdog idea, and have put that in. I have staggered the
report 1s as you suggest. And changed the retrys down to 2.

I am still using v5.0 of the 11_FHT.pm

thanks again!

Al

[Zrrronggg!]

> I like the watchdog idea, and have put that in.

I got that idea from somebody else here on the group and it works very
well for me.

> I am still using v5.0 of the 11_FHT.pm

There still might be another problem, but I have no idea what. I only
saw after reading your conf. that you might be in trouble anyhow. Lets
us know if the changes work for you.

> thanks again!

NP