Closed GoogleCodeExporter closed 8 years ago
String ZwaveNode01HomeID "Home ID [%s]" (gZwaveNode01)
{zwave="1:1:command=info,item=home_id"}
Number ZwaveNode01NodeID "Node ID [%s]" (gZwaveNode01)
{zwave="1:1:command=info,item=node_id"}
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:04
I think so ben regarding dead, I will fix this.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:05
I've fixed it. It's in my clone. I will provide binaries shortly. I've fixed
some other bugs as well.
I started working on saving and restoring of the configuration state as well.
For now i've only implemented saving using XStream as XML serializer. Some
example xmls are included below. Once I provide binaries I'm very interested in
your XML's as well. They are saved in the etc/zwave directory.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:16
Attachments:
I have been keeping up-to-date with your 1.4.0 fork - if I get the latest it
should be stable? I am happy to build locally.
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 8:18
oh man...I didn't use the underscore. Thanks. Works now.
For the siren I have to have a closer look.
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 8:22
Pretty stable... Although I want to run it overnight to be sure ;-)
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:22
I am willing to risk it ;). Will post my node files here once they are
generated.
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 8:24
I've made them interchangeable so that we can start sharing those in the same
way as open-zwave does. I'll put in some checks on restoring so that invalid
versions (both of the binding and node firmware) are not mixed.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:29
And a binary for 1.3.1
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:48
:)
Could you please add it? :D
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 8:51
DerRocka, I don't see any switch_binary_set messages in your log for node 3.
Are you sure you're actually toggling the switch? Can you provide a complete
log (not just Z-Wave).
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:52
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:52
Attachments:
Hmmm... Gotta find another way of distributing those. I'm out of space... quota
exceeded.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:53
Ok - node XML files being generated. They are all there apart from my door
sensor (FGK101) as that only wakes up every hour. I will wait till that is
generated and then post a full set. Looking good so far ;).
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 8:55
Dropbox share?
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 8:57
Might even taken two cycles so two hours :-)
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 8:57
Oh okay.
yes, of course. Well...I could quota exceeded. hmm...no idea right now.
I am toggling the switch item in HABdroid. But I have no rule or something like
that. I thought that could be done with the item-definition itself. Like the
other switches.
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 8:59
Yes, that should work but the event does not get to the binding somehow. It
cannot send what it doesn't receive :-)
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 9:04
True. But what could be the reason of that?
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 9:06
Cannot tell without a complete log. I do see node 2 toggled from habdroid, but
that's probably another real switch.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 9:21
yes it is another switch. can't find a switch_binary_set either. But can't
upload at the moment.
In my event.log I can see my item:
2013-10-01 23:27:16 - Alarm_Sirene received command ON
2013-10-01 23:27:25 - Alarm_Sirene received command OFF
At this time I should see something in openhab.log or in zwave.log shouldn't I?
Nothing to see.
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 9:33
Could you post your items file?
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 9:36
Hmmm - think it might be a 'quota exceeded' for this issue, rather than us as
users. I can't attach my node files either. I have emailed them to you directly
JwS.
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 9:39
Yes:
Switch Alarm_Sirene "Alarm-Sirene" <siren-on>
(gAlarmSystem) { zwave="3:command=switch_binary" }
Number Alarm_Sirene_Battery "Alarm-Sirene Batterie [%s]"
(gZwaveStats) { zwave="3:command=battery" }
Switch Tuer_Sensor "Tür Sensor [%s]"
(gAlarmSystem) { zwave="4:command=basic" }
//String Tuer_Sensor_Basic "Tür Sensor [%s]"
(gAlarmSystem) { zwave="4:1:basic" }
Number Tuer_Sensor_Battery "Tür Sensor Batterie [%s]"
(gZwaveStats) { zwave="4:command=battery" }
/* Zwave Controller stick statistics */
Number ZwaveStatsSOF "Number Start of Frames[%s]"
(gZwaveStats) { zwave="1:command=info,item=sof" }
Number ZwaveStatsACK "Number of ACK [%s]"
(gZwaveStats) { zwave="1:command=info,item=ack" }
Number ZwaveStatsCAN "Number of CAN [%s]"
(gZwaveStats) { zwave="1:command=info,item=can" }
Number ZwaveStatsNAK "Number of NAK [%s]"
(gZwaveStats) { zwave="1:command=info,item=nak" }
Number ZwaveStatsOOF "Number of OOF [%s]"
(gZwaveStats) { zwave="1:command=info,item=oof" }
.
.
.
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 9:39
Crap Google.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 9:40
I'll create a new issue tomorrow with an updated todo. Now on a tablet.
Original comment by jwsp...@gmail.com
on 1 Oct 2013 at 9:42
Good idea. You get those node files ok?
Original comment by ben.jone...@gmail.com
on 1 Oct 2013 at 9:43
Yes. And it's getting late. So now I couldn't test it anyway :)
Original comment by DerRo...@gmail.com
on 1 Oct 2013 at 9:44
I've deleted attachments and it's still exceeded. jeez.
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 8:10
Like here:
https://code.google.com/p/support/issues/detail?id=30796
http://code.google.com/p/support/issues/detail?id=28358
A quota can be increased by "Project members"?
Original comment by DerRo...@gmail.com
on 2 Oct 2013 at 8:19
Man man...if you have a closer look at your line and my line of the zwave
binding for the siren you can see, the only difference "<siren-on>". Without it
everything works fine and I can hear the 100db too :)
Original comment by DerRo...@gmail.com
on 2 Oct 2013 at 9:01
Great to hear.
BTW, I've moved the development issue list to my clone at github:
https://github.com/jspuij/openhab/issues?state=open
This is only for development issues. I'll update the wiki page to reflect this.
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 9:10
I'll provide binaries for my 1.4.0-zwave branch using cloudbees from now on:
https://jspuij.ci.cloudbees.com/job/openhab-zwave/
Will add this to the wiki as well
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 9:26
If I try to connect to cloudbees the newly created github authorization fails.
Original comment by DerRo...@gmail.com
on 2 Oct 2013 at 10:54
Hmmm.. It should be public, without authorization. Hold on.
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 11:05
I made it publicly accessible and now it's gone... long live cloudbees.
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 11:15
Ah apparently it takes a little time before everything is processed. It should
be accessible now.
Original comment by jwsp...@gmail.com
on 2 Oct 2013 at 11:18
Yes it works.
Original comment by DerRo...@gmail.com
on 2 Oct 2013 at 11:29
I have asked for having the quota increased:
https://code.google.com/p/support/issues/detail?id=31068
Original comment by kai.openhab
on 4 Oct 2013 at 6:34
Ben, All.. I've finished configuration serialization
https://github.com/jspuij/openhab/issues/5. Starting the binding saves me about
2 minutes ;-)
Original comment by jwsp...@gmail.com
on 4 Oct 2013 at 12:49
BTW, Thanks Kai...
Original comment by jwsp...@gmail.com
on 4 Oct 2013 at 12:49
Nice one JwS. Is it in your clone? Do I need to delete those existing config
files I generated for you the other day or is the format unchanged? I will
pull, compile and test ASAP!
Original comment by ben.jone...@gmail.com
on 4 Oct 2013 at 7:26
Format is the same, so you should be good to go immediately.
Original comment by jwsp...@gmail.com
on 4 Oct 2013 at 7:54
Seems to be working nicely. Great work again! Just need to wait for my door
sensor to finally initialise and serialise itself, and then the whole network
should come up instantly on a restart. Noticed you create a sub-directory for
the binding version #. Probably quite sensible going forward - I presume the
idea is when I build a new version of the binding I can just copy the previous
version xml files into the new directory and it will save having to rebuild
them?
Original comment by ben.jone...@gmail.com
on 4 Oct 2013 at 10:50
Got a quick question JwS - I am seeing a lot of 'noise' in my openHAB event log
due to all the ZWave info updates. I have cut my info bindings right back to
just the network stats and the NodeDead flag, but every 10s I still see about
30 event updates. In order to reduce the frequency of these updates I presume I
can just increase the 'refresh timeout' in openhab.cfg for the ZWave binding?
And since I don't have polling turned on for any nodes this won't effect
anything else? Just wondering if this info polling should have a separate
config value - since in reality would you ever want this stuff to be updated so
frequently? I was thinking once every 10mins would be enough - or maybe even an
hour? Or can I just do this by setting the polling interval for each individual
'info' item binding?
Not a big issue - just something I thought I would note down while I remembered!
Original comment by ben.jone...@gmail.com
on 4 Oct 2013 at 11:05
Yeah. I was thinking about fixing that. I don't like the current structure with
the refresh in it and the AbstractActiveBinding inheritance. I was thinking
about removing it and doing everything with events. For now increase the
refresh time, or you can add a specific binding refresh time to an info binding
as well I think...
Original comment by jwsp...@gmail.com
on 7 Oct 2013 at 6:47
JwS - got something a little weird going on - nothing major but I am unable to
track down what is causing it so thought you might be able to help. Attached is
a log showing the issue.
If you search for 'ERROR' you should find 5 instances. You can ignore the
first, that is just my door sensor being asleep.
The other 4 occur at two instances. Both when I am turning off all my lights -
the first when I went to bed at 10:59pm and the second at 07:31am when the sun
rises. I have simple rules for turning off all lights at these times.
And it is always Node 3 - which happens to be the only AeoTec Smart Energy
Micro Switch device I have.
I have tried manually turning this switch on/off both at the wall, and via
openHAB, and I can't reproduce the error. It only seems to happen when it is
doing the 'all lights off' rule.
What causes the controller to cancel a message? Is it the device itself
bouncing it back? Or is the controller getting overloaded with too many things
to do at once? Seems a bit odd that it is ALWAYS node 3, but maybe that is just
because of the order of execution?
Any ideas? It is not a major issue, as everything seems to be working fine
regardless, I am just keeping an eye on the ZWave stats however and seeing this
'cancelled' count increasing each day is bugging me ;).
Original comment by ben.jone...@gmail.com
on 9 Oct 2013 at 9:37
PS - I will email you the log since I can't attach it here anymore.
Original comment by ben.jone...@gmail.com
on 9 Oct 2013 at 9:37
Disclaimer: have not looked at the log yet.
It's probably nothing. Cancels (and retransmissions) occur when the
controller both receives a message from our binding, and is also sending a
message from the network to the binding. Although the serial communications
channel would certainly allow for it, the Z-Wave controller is only half
duplex.
Your energy meter will probably report a change in the power consumption
immediately after switching the device on or off... This message is received
by the binding while it it's flushing its queue with "off" messages to all
nodes. For the binding it's not a problem, but the controller cannot process
the next "off" command for another node we send, while its also sending us
the new power level of the previous node. Hence the CAN.
We receive the CAN and resend the canceled message by the controller. I have
exactly the same issue with my Fibaro WALL switch and my all off sequence.
Nothing to worry about and nothing we can/should do about it. You never know
when an unsollicited message from the network arrives. It can always happen
while we send a message.
JwS
Original comment by jwsp...@gmail.com
on 10 Oct 2013 at 11:19
We could lower the error level to have it disappear from the logs, but in case
there are many it indicates an error in our own transaction handling. We have a
transaction semaphore that guards that a new message is not sent before the
previous sequence of messages is finished. This way we can guard that we don't
create CAN messages ourselves.
This does not protect us however from unsollicited messages from one of the
nodes in the network.
Original comment by jwsp...@gmail.com
on 10 Oct 2013 at 11:33
Original issue reported on code.google.com by
jwsp...@gmail.com
on 2 Sep 2013 at 12:45