Closed GoogleCodeExporter closed 8 years ago
Hi,
is it possible to implement a Node-Count or am I just missing something?
Original comment by DerRo...@gmail.com
on 3 Sep 2013 at 12:04
Hi there,
I was following your development of the ZWave-binding and have to say: Great
job, guys!
Anyhow I was wondering where the METER_CLASS went. Ben made one, mentioning it
in issue #265, but apparently it never made it to the default branch. So far it
would be alright, but I can't even find it in this tasklist you made. Did you
skip it for a reason or were it just forgotten?
Thanks for your attention.
Original comment by der.tim....@gmail.com
on 3 Sep 2013 at 4:43
METER will be merged from Ben into the 1.4.0 version of the binding. I will
probably release a version of the 1.4.0 binding next week. Have been a bit busy
with the 1.3 release.
Original comment by jwsp...@gmail.com
on 3 Sep 2013 at 7:04
Der Rocka,
You mean like a report item? Yes, that would very well be possibly. Never
thought of it ;-) I will add it to the list. Have to find a better way to
manage these items as I can't seem to edit comments...
Original comment by jwsp...@gmail.com
on 3 Sep 2013 at 7:07
Hi,
yes like a report item.
I just noticed it in the OZWCP and thought that it could be quite usefull. Just
to see how many nodes my network does have in general. Perhaps also for new
added nodes, that I can see, that they were added right. At least that there
was added one node before I can try to add it to my items-file. :-)
Regards
André
Original comment by DerRo...@gmail.com
on 4 Sep 2013 at 7:40
André,
I think you have an Aeon labs controller that is able to include devices on
it's own. I'm not sure which messages are sent (eg. programming return routes
in a node, requesting neighbor updates) by the stick in this inclusion process,
so it could be that a node is somewhat unusable after inclusion by the stick.
Then you'd still need OZWCP to run an initialization before all is well in the
network. The binding currently does not assign return routes or requests nodes
to update their neighbor list as part of it's initialization phase.
Original comment by jwsp...@gmail.com
on 4 Sep 2013 at 10:18
Jan-Willem,
you're right. I don't know exactly whether there is a usefull thing about a
Node-Count for everyone, but I came across in OZWCP, that I find this
information nice.
It isn't really much about the thing, that I can see whether a node was
successfully added or not. Just to see how many a network has in whole. In my
little network at the time (with just 1 node) there isn't really a use for it,
but perhaps nice to have ;-)
André
Original comment by DerRo...@gmail.com
on 4 Sep 2013 at 11:04
Another Website with Z-Wave Productinformation by sigmadesigns. Not like the
other, I gave you before, but with official information by vendors etc:
http://products.z-wavealliance.org/
There you can download a pdf-file for every official supported product with
their supported Device/Command/Controller Classes. Perhaps also worth to be on
the wikipage?
In special for example the Fibaro Double Relay Switch (It's an 1 page pdf):
http://products.z-wavealliance.org/products/46
André
Original comment by DerRo...@gmail.com
on 4 Sep 2013 at 11:16
Thanks, I've added it to the wiki page.
Original comment by jwsp...@gmail.com
on 4 Sep 2013 at 11:40
Little heads up. I've laid the ground work for the converter stuff. Hopefully I
can finish it by the end of the week.
Original comment by jwsp...@gmail.com
on 4 Sep 2013 at 9:22
Hi everybody,
First of all, thanks for this great contribution.
I've been following the zWave thread, and I have just downloaded the 1.3
snapshot binding. However it does not work as I expected.
I have an Aeon Labs USB Z-Stick Zwave controller, and a Fibaro Controlled
Switch (Fibaro calls it FGWPE/F-101). It also can be used to measure power
consumption.
Both devices work fine when used outside openHAB. I tested with Zwave.me and I
have been able to control the switch and to query its state.
Fibaro switch has Id=2 and the network is functioning, as the Zwave.me tests
reveal.
However, when I use them with openHAB, the Fibaro switch behaves kind of
weird...
OpenHAB acknowledges when I turn the switch on/off (using the manual button on
the Fibaro switch) because the web interface shows the appropiate state, but I
cannot make openHAB turn the switch on/off FROM the web interface.
It seems that I can read the satus of the device but I cannot send commands to
it.
Besides, openHAB always says that device is dead when I query the Sleeping_Dead
state (both when the device is connected and disconnected from the wall)
What am I doing wrong???
What I need is:
-to be able to control the switch with OpenHAB web interface
-to know if the switch has been swichted off/on
-to know if the switch has lost mains power supply (it is dead)
My items file contains these lines:
String ZwaveNode02Id "Node ID [%s]" {zwave="2:1:NODEID"}
Switch ZwaveNode02Switch "Fibaro Switch" {zwave="2"}
String ZwaveNode02Dead "Node Is Dead [%s]" {zwave="2:1:SLEEPING_DEAD"}
And my sitemap contains these lines:
Frame label="Zwave"
{
Text item=ZwaveNode02Id
Switch item=ZwaveNode02Switch icon="light"
Text item=ZwaveNode02Dead icon="light"
}
The Fibaro switch is in front of the Zwave controller so I am sure it is not a
range problem.
Any ideas??
Original comment by serca...@gmail.com
on 7 Sep 2013 at 5:59
This is a bit of a problem device. I have one myself. It has a firmware bug
that prevents it from initializing properly and it has an energy meter combined
as well in a way that the binding currently does not support. The next beta
release, which is imminent, will support it however.
JwS
Verstuurd vanaf mijn iPhone
Original comment by jwsp...@gmail.com
on 7 Sep 2013 at 7:03
Thanks Jan-Willem.
I was starting to think that I was a fool :-)
Does Fibaro know the bug?
Is there a firmware upgrade to solve it?
In case they don't know, I think we should tell them, don't you think?
Thanks JW!
Original comment by serca...@gmail.com
on 7 Sep 2013 at 8:44
Thanks Jan-Willem.
I was starting to think that I was a fool :-)
Does Fibaro know the bug?
Is there a firmware upgrade to solve it?
In case they don't know, I think we should tell them, don't you think?
Thanks JW!
Original comment by serca...@gmail.com
on 7 Sep 2013 at 8:44
Hey JW - back from holidays and feeling refreshed! Looks like you have been
pretty busy. I downloaded the 1.3.0 release and ran up the ZWave binding but
had to make a couple of modifications to get my METER node working and had to
change the BINARY_SENSOR and BASIC command classes to return OPEN/CLOSED rather
than ON/OFF as that is how my config is setup.
Anyways, I have been noticing a lot of errors. These were happening before
1.3.0 final, and are still happening. All the nodes seem to respond ok although
it does sometimes take a couple of openHAB restarts to get them all to come up.
Some responses seem to get missed and the node gets stuck in VERSION or
MANUFACTURER, it can vary.
I am guessing it is the same issue I am seeing once the binding is running, but
when it is running it just ignores those failures since it is all initialised
already.
Attached are my logs. The send/receive threads are dying a lot which is a bit
of a worry. I have removed the USB hub and now have the stick connected via a
500mm USB extension cord, as I know you mentioned that could be a cause of
serial problems.
There are also a lot CAN messages from the controller. I am starting to wonder
if there is not something wrong with my USB stick?!
Anyways - let me know if you see anything obvious. Not sure what else I can try.
Cheers,
Ben
Original comment by ben.jone...@gmail.com
on 9 Sep 2013 at 7:16
Attachments:
Hmmm - after looking at that log some more it looks like the send/receive
threads always seem to die just after a METER_EVENT is received. That is a bit
of a red flag! I will look into that tomorrow as this is my code!
Original comment by ben.jone...@gmail.com
on 9 Sep 2013 at 7:22
Sercasyr,
Fibaro has been notified and they promised a fix. We will create a workaround
though, this is way easier. Normally Fibaro puts the firmware on new devices
and they make them available through their own gateway product, Homecenter 2.
I've never seen separate firmware updates for devices.
Original comment by jwsp...@gmail.com
on 9 Sep 2013 at 7:33
Ben, welcome back, let me know what you find.
Original comment by jwsp...@gmail.com
on 9 Sep 2013 at 7:38
Ok - well I couldn't see anything obviously wrong, however after disabling
support for the METER command class (0x32) (by setting it null in the
ZWaveCommandClass.CommandClass enum) and removing the item from my config, the
binding is humming along perfectly. No errors, timeouts, or thread restarts.
The only error I see is the 'Unsupported command class METER' whenever my
switch/meter node sends a METER report.
I have no idea what that command class is doing to cause such mayhem but it is
a little concerning! Has anyone else managed to get METER reports working on
any other devices?
Original comment by ben.jone...@gmail.com
on 9 Sep 2013 at 9:10
It looks like the switch only sends meter reports every 10 mins, but my binding
was failing every minute (at least). So I am thinking this problem is to do
with the binding polling the node and requesting a meter report, via the
METER_VALUE binding action, which happens every 60 seconds. But I am still
unsure exactly what that problem is...
Original comment by ben.jone...@gmail.com
on 9 Sep 2013 at 9:20
I'll have a look at it tomorrow. I'm pretty much ready with 1.4.0 alpha to
incorporate meter.
JwS
Original comment by jwsp...@gmail.com
on 9 Sep 2013 at 10:24
Nice one - I have been keeping an eye on your 1.4.0 branch - looks great. Let
me know when it is ready for testing and I will start using it here.
Original comment by ben.jone...@gmail.com
on 9 Sep 2013 at 11:11
OK, I've completed the items below in my 1.4 branch.
+ D make binding between command classes and items more dynamic. So that
switches can be bound to dimmers, sensors to contacts or switches, etc.
+ M initialize / read binding values in the DYNAMIC node stage during binding
initialization.
+ M Make polling a binding action, so that it is configurable from the items
file. Default should be no polling.
+ M Make binding actions for alarm types, so that binding can be done on a
specific alarm. Remove alarm values from the binding again then, to make it
stateless.
Original comment by jwsp...@gmail.com
on 10 Sep 2013 at 11:23
And I merged the METER class. Thanks Ben. I'm happy to announce that the
binding now fully supports the Fibaro FGWPE/F-101 switch witch included power
and energy meters. I have one myself (connected to my espresso machine) and it
works beautifully.
I'll test it for a few days with my setup because I've changed about every bit
of code in the binding, but will provide binaries before the end of the week.
For those of you interested, the configuration part for the Fibaro FGWPE/F-101
is below:
Switch Coffee_Kitchen_Switch "Koffiemachine" (GF_Cellar)
{zwave="18:command=switch_binary"}
Number Coffee_Kitchen_Power "Koffiemachine actueel verbruik [%.1f W]"
(GF_Cellar) { zwave="18:command=sensor_multilevel" }
Number Coffee_Kitchen_Energy "Koffiemachine totaal verbruik [%.2f KWh]"
(GF_Cellar) { zwave="18:command=meter" }
What is immediately obvious is that a command class is specified (besides the
node ID) because this device has only one endpoint, but multiple different
command classes.
There have been more changes to the binding format, but in general the binding
format is: (brackets indicate optional values)
<node id>[:<endpoint id>][:<argument1>=<value1>[,<argument2>=<value2>]]
Some more examples are:
String ZwaveStatsSOF "Number Start of Frames[%s]" (gZwaveStats)
{zwave="1:command=info,item=sof"}
String ZwaveStatsACK "Number of Acknowledgments [%s]" (gZwaveStats)
{zwave="1:command=info,item=ack"}
String ZwaveStatsCAN "Number of CAN [%s]" (gZwaveStats)
{zwave="1:command=info,item=can"}
Contact Door_Living_Switch "Schuifdeur sensor [MAP(nl.map):%s]" (GF_Living)
{zwave="25:command=sensor_binary"}
Number Door_Living_Sensor_Battery "Schuifdeur sensor batterijniveau [%d %%]"
(GF_Living) { zwave="25:command=battery" }
Dimmer Light_Toilet_Dimmer "Toilet Dimmer [%d %%]" (GF_Toilet)
{zwave="10:1:restore_last_value=true"}
Switch Mech_Vent "Mechanische ventilatie middel." (GF_Kitchen) {zwave="11:1"}
Switch Mech_Vent_High "Mechanische ventilatie hoog." (GF_Kitchen)
{zwave="11:2"}
Original comment by jwsp...@gmail.com
on 10 Sep 2013 at 6:39
Looks awesome JwS! Can't wait to get my hands on this. My binding seems to be
working a lot more reliably now I am not polling for METER values. Did you make
any changes to my METER command class - find any bugs? I presume it is working
ok for you?
I am not too fussed as I only have one node which reports METER values, and
that is on a single LED outdoor light so uses next to no power at all.
Original comment by ben.jone...@gmail.com
on 10 Sep 2013 at 8:53
Built your latest version (1.4.0) of the binding and have been testing it out
today. I definitely have something funny going on with my Aeon Labs Micro Smart
Energy Switch. I have disabled meter reports (I think!) but when the binding
starts up it is requesting a value during initialisation.
It seems like anytime this switch sends a meter report the message is received
ok but then something (I presume the ACK?) times out and the send/receive
threads die. Logs attached. So I really don't want these meter reports being
sent! In 1.3.0 I just removed the item binding for the METER command and
everything was sweet (the node is right next to the controller so definitely
not a network issue).
So can you confirm the meter report request is sent if a node reports support
for the METER command class? And if so, do you think we can disable this if
there is no item binding for this action, since there is nothing to initialise?
Got a few other things but I will try and keep each message related to a single
issue.
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 6:24
Attachments:
I'm pretty sure that the METER report is delivered to an incompatible item.. a
switch or dimmer or something and that an exception is thrown.
Could you send me the binding configuration for this item? try explicitly
specifying the command class (the other one, not METER) in the binding config.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 8:57
I think I fixed it. I'm now also considering the type of the z-wave value when
picking a converter.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 9:21
The thing is I didn't have any item configured for the meter value, all I have
is a simple Switch item for the node {zwave="3"} - i.e. it is just a light
switch. Or are you suggesting I try something like
{zwave="3:command=SWITCH_BINARY"} to ensure the meter report is ignored?
I will try your new code tomorrow first thing - it is getting a little late
here to be firing up the dev PC and compiling binaries!
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:31
BTW - what is the syntax for enabling polling for an item? All my Fibaro
switches are pretty good at reporting state changes, but the Aeon Labs Micro
Switch doesn't seem to do it so I will need to poll that one node.
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:32
refresh_interval=xx where xx is the amount of seconds.
and yes, I mean {zwave="3:command=SWITCH_BINARY"}, this will ignore the meter
reports.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 10:41
And try associating the Aeon Labs switch with the controller in open-Zwave. It
will probably start reporting it's values.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 10:43
I did have it associated and at one point it was reporting its values, but
since the meter issue I was trying to stop any meter reports so removed all
associations. Once I get the node working ok (i.e. not crashing whenever a
meter report arrives!) I will try and get the associations working again.
I did notice with my other sensors that they send BASIC reports when they are
tripped (i.e. Door Sensor and the Universal Sensor). I had Contact items
configured with binding configs like {zwave="10:command=SENSOR_BINARY"} but my
items were not being updated when the sensor was tripped.
Is this the correct format? I tried command=BASIC as well, but this gave a
warning about a missing converter type or something (sorry don't have the log
in front of me). Can you see anything wrong with my initial attempt using
SENSOR_BINARY, knowing that only BASIC reports are coming back?
Is this something wrong with the sensor config, or should your new converter
stuff handle this case and convert the BASIC report to an Open/Closed contact
value?
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:50
And I did not really change the METER command class. I've expanded it a bit so
that it supports up to version 3 of the command class and I've added RESET
functionality, that you can bind to an OnOff item (and create a push button in
the sitemap) to reset the energy meter. That looked like a nice gimmick.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 10:51
Yeah I saw that reset stuff. Quite useful I guess - I would love to get my
hands on one of your wall plugs but they are not available for AUSNZ yet...
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:53
OW yeah, I forgot about the weird BASIC reports from the door sensors...
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 10:57
I can't remember what you did exactly in 1.3.0 to fix these. Something along
the lines of handling BASIC reports as a 'fallback'? The problem I had was that
initially you were sending ON/OFFs for the BASIC reports - but I wanted to use
Contact items - so I had a custom version of your binding which handled them as
OPEN/CLOSED.
I was hoping all your new converter stuff would eliminate the need for my
customisation.
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 11:11
I completely forgot the basic converter. Oops. Creating it now.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 2:29
I've fixed it, it's in my clone. I completely looked over it, had a few non
functioning sensors myself. Anyway, my binding strings to the door sensor:
Contact Door_Living_Switch "Schuifdeur sensor [MAP(nl.map):%s]" (GF_Living)
{zwave="25:command=sensor_binary,respond_to_basic=true"}
Number Door_Living_Sensor_Battery "Schuifdeur sensor batterijniveau [%d %%]"
(GF_Living) { zwave="25:command=battery" }
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 3:21
Hmm Ben,
Meter seems more complicated than we thought. I have a Greenwave powerbar
(actually Schuko) that has meter support. It seems that it can measure both
power in W and energy in KWh. But instead of a separate command class
SENSOR_MULTILEVEL (like the fibaro), it uses the scales feature to report them
as different scales.
I'll adjust it accordingly.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 5:26
Hi JWSP!
I am a little bit confused - is this new binding config format (command=...)
supported only in future 1.4.0 version, or in current 1.3.0 too?
Original comment by roher.ro...@gmail.com
on 11 Sep 2013 at 5:31
Only in 1.4.0 i'm afraid. This thread is about the next version of the binding.
I'll provide some alpha binaries, but as you see, stuff breaks here.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 6:20
JW - I can't see anything in your clone since the fix from comment #28.
BTW - I did notice there was no basic converter but figured you had some other
cunning way of handling those messages - my bad - I should have alerted you!
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 8:21
forgot to push. Sorry.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 8:42
Great - thanks! So I have compiled your changes and things are looking better.
The METER reports are not arriving anymore, so node 3 is correctly
initialising. And the BASIC reports from my sensors are being handled - many
thanks!
Next problem however. Node 4 is failing to initialise. This is a standard
Fibaro 2x1.5kW switch. Below are the log entries relating to it during
initialisation. It looks like the binding receives an application command
request and then dies. After this the binding is unable to recover and the node
is marked dead.
This looks like another one of my timeout issues, and node 4 is quite far from
the controller, but I wouldn't expect a timeout to kill the send/receive
threads. If there are exceptions being thrown, is there any way we can dump the
details to the logs?
2013-09-12 09:09:43 DEBUG
o.o.b.z.i.p.ZWaveController$ZWaveReceiveThread[:1133]- Message = 01 05 00 04 08
04 F2
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.SerialMessage[:146]- Creating new
SerialMessage from buffer = 01 05 00 04 08 04 F2
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.SerialMessage[:150]- Message checksum
calculated = 0xF2, received = 0xF2
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.SerialMessage[:163]- Message Node ID = 255
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.SerialMessage[:164]- Message payload = 08
04
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.ZWaveController[:145]- Incoming message
to process
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.SerialMessage[:238]- Assembled message
buffer = 01 05 00 04 08 04 F2
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.ZWaveController[:146]- Message: class =
ApplicationCommandHandler (0x04), type = Request (0x00), buffer = 01 05 00 04
08 04 F2
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.ZWaveController[:166]- Message type =
REQUEST
2013-09-12 09:09:43 DEBUG o.o.b.z.i.p.ZWaveController[:192]- Application
Command Request from Node 4
2013-09-12 09:09:48 ERROR o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:1019]-
Timeout while sending message to node 4. Requeueing
2013-09-12 09:09:48 ERROR o.o.b.z.i.p.ZWaveController[:308]- Got an error while
sending data to node 4. Resending message.
2013-09-12 09:09:48 DEBUG o.o.b.z.i.p.ZWaveController[:847]- Callback ID = 15
2013-09-12 09:09:48 DEBUG o.o.b.z.i.p.ZWaveController[:717]- Enqueueing
message. Queue length = 7
2013-09-12 09:09:48 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:996]-
Took message from queue for sending. Queue length = 6
2013-09-12 09:09:48 DEBUG o.o.b.z.i.p.SerialMessage[:238]- Assembled message
buffer = 01 08 00 13 04 01 00 25 0F CB
2013-09-12 09:09:48 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:1007]-
Sending Message = 01 08 00 13 04 01 00 25 0F CB
2013-09-12 09:09:52 DEBUG o.o.b.z.i.ZWaveActiveBinding[:98]- Zwave Network
isn't ready yet!
2013-09-12 09:09:52 WARN o.o.b.z.i.p.ZWaveController$WatchDogTimerTask[:1199]-
Threads not alive, respawning
2013-09-12 09:09:52 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:1036]-
Stopped Z-Wave send thread
2013-09-12 09:09:52 INFO o.o.b.z.i.p.ZWaveController[:708]- Disconnected from
serial port
2013-09-12 09:09:52 INFO o.o.b.z.i.p.ZWaveController[:627]- Connecting to
serial port /dev/zwave
2013-09-12 09:09:52 DEBUG
o.o.b.z.i.p.ZWaveController$ZWaveReceiveThread[:1090]- Starting Z-Wave receive
thread
2013-09-12 09:09:52 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:991]-
Starting Z-Wave send thread
2013-09-12 09:09:52 INFO o.o.b.z.i.p.ZWaveController[:640]- Serial port is
initialized
2013-09-12 09:09:52 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:996]-
Took message from queue for sending. Queue length = 5
2013-09-12 09:09:52 DEBUG o.o.b.z.i.p.SerialMessage[:238]- Assembled message
buffer = 01 08 00 13 05 01 00 25 04 C1
2013-09-12 09:09:52 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:1007]-
Sending Message = 01 08 00 13 05 01 00 25 04 C1
2013-09-12 09:09:52 ERROR
o.o.b.z.i.p.ZWaveController$ZWaveReceiveThread[:1148]- Message cancelled by
controller (CAN), resending
2013-09-12 09:09:53 DEBUG o.o.b.z.i.p.ZWaveController[:717]- Enqueueing
message. Queue length = 6
2013-09-12 09:09:53 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:996]-
Took message from queue for sending. Queue length = 5
2013-09-12 09:09:53 DEBUG o.o.b.z.i.p.SerialMessage[:238]- Assembled message
buffer = 01 08 00 13 05 01 00 25 04 C1
2013-09-12 09:09:53 DEBUG o.o.b.z.i.p.ZWaveController$ZWaveSendThread[:1007]-
Sending Message = 01 08 00 13 05 01 00 25 04 C1
2013-09-12 09:09:53 ERROR
o.o.b.z.i.p.ZWaveController$ZWaveReceiveThread[:1148]- Message cancelled by
controller (CAN), resending
2013-09-12 09:09:53 DEBUG o.o.b.z.i.p.ZWaveController[:717]- Enqueueing
message. Queue length = 6
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:12
In another attempt to initialise I got a version = 0 when initialising a multi
instance node. This caused the binding to report a warning but then it couldn't
proceed and the node was marked dead. Looking at your code would it be possible
to trap this error and attempt to retry this phase of initialisation (i.e.
resend the request)?
Original comment by ben.jone...@gmail.com
on 11 Sep 2013 at 10:31
No idea about logging, usually I run the binding in the debugger during
development. The binding should not throw exceptions on timeouts. When you have
logged one, I'll look into it.
Original comment by jwsp...@gmail.com
on 11 Sep 2013 at 11:01
I can't run mine in the debugger unfortunately, my dev pc is too far away from
the controller, and I am loathed to move it for fear of messing up my mesh
network.
Now, after a few restarts, I have all nodes up and running. I will keep an eye
on things and continue to dig around your code to see if I can see what is
going wrong (when it does).
Cheers.
Original comment by ben.jone...@gmail.com
on 12 Sep 2013 at 12:16
JwS - hats off to you my friend. This latest version of the binding is
fantastic. I have managed to get all my config tweaked and now all my nodes
happily report state change events when manually switched. There is next to no
traffic on my ZWave network, except for the occasional temperature report and
of course when something is switched/activate.
The new converter stuff seems to be working great. I have switches and contacts
receiving BASIC reports and they all update seamlessly.
One thing I did think, would it be worth making the RESPOND_TO_BASIC command ON
by default? I have a couple of nodes/sensors that send BASIC reports and I
would be surpised if there weren't more out there. Rather than force a user to
explicitly set this option, it might save a few headaches if it was enabled by
default. I can't think of any reasons you wouldn't want to handle a BASIC
report, but you can always disable the option explicitly if required.
Just a thought - otherwise everything seems to be humming along swimmingly!
About time to get some more devices I think...
Original comment by ben.jone...@gmail.com
on 12 Sep 2013 at 4:08
I'm glad it works for you :-) Yeah, I think I can make the binding part a bit
more clever. Basic can be enabled by default, but then I have to make a best
guess which item is the "basic" item. i.e. with the door sensor, there is also
an item bound to the battery command class, and we don't want that one to
respond to the basic reports.
Maybe it's also time to start a wiki page / DB with binding config lines per
device. I think maybe Thomas was right the other day, it will help novice users
to quickly set-up a their network.
Holy grail of course is dynamic creation of items based on the discovered
nodes. This way, there is no configuration anymore...
Original comment by jwsp...@gmail.com
on 12 Sep 2013 at 6:21
Original issue reported on code.google.com by
jwsp...@gmail.com
on 2 Sep 2013 at 12:45