Starting latest Zwave crash OH2

[marcel_verpaalen (Marcel)]

Hi Cris,

Maybe you have a hint on what is going wrong.
I pulled latest OH2 from cloudbees, using the included zwave jar. During the startup OH2 completely crashes (even no error in the screen, full back to command line) after the initialization of the /dev/ttyUSB0.

OH1 is running fine with this Aeon Z-wave S2 stick. When removing the port= /dev/ttyUSB0 OH2 runs fine incl habmin. Hence I don’t expect rights issue

2015-08-30 22:53:08 [DEBUG] [e.s.i.t.upnp.UpnpIOServiceImpl:152  ] - Starting UPnP IO service...
2015-08-30 22:53:09 [DEBUG] [.z.internal.ZWaveActiveBinding:175  ] - ConverterHandler not initialised. Polling disabled.
2015-08-30 22:53:09 [INFO ] [.z.internal.ZWaveActiveBinding:325  ] - Update config, port = /dev/ttyUSB0
2015-08-30 22:53:09 [INFO ] [.z.internal.ZWaveActiveBinding:330  ] - Update config, healtime = 4
2015-08-30 22:53:09 [INFO ] [.z.internal.ZWaveActiveBinding:375  ] - Update config, setSUC = true
2015-08-30 22:53:09 [INFO ] [.z.internal.ZWaveActiveBinding:384  ] - Update config, softReset = true
2015-08-30 22:53:09 [INFO ] [.z.internal.ZWaveActiveBinding:393  ] - Update config, masterController = true
2015-08-30 22:53:09 [DEBUG] [.z.internal.ZWaveActiveBinding:282  ] - Initialising zwave binding
2015-08-30 22:53:09 [INFO ] [.service.AbstractActiveService:169  ] - ZWave Refresh Service has been started
2015-08-30 22:53:09 [INFO ] [b.z.i.protocol.ZWaveController:146  ] - Starting Z-Wave controller
2015-08-30 22:53:09 [INFO ] [b.z.i.protocol.ZWaveController:154  ] - Z-Wave timeout is set to 5000ms. Soft reset is true.
2015-08-30 22:53:09 [INFO ] [b.z.i.protocol.ZWaveController:320  ] - Connecting to serial port /dev/ttyUSB0
m

[chris (Chris Jackson)]

Hi Marcel,
I just compiled the binding myself, and it seems to be ok here (again, OH1 binding running under OH2). It seems to be working ok here…

So, either it’s the Cloudbees compilation (it’s not unheard of for it to go wrong), or there’s something specific about your configuration. I think it’s unlikely it’s just your config as it’s not getting far enough to talk to devices etc…

I’m not sure that I can attach the version I’ve compiled to the new forum, so maybe try tomorrows version from Cloudbees, or email me and I can send it to you to try…

Chris

[marcel_verpaalen (Marcel)]

Thanks Chris,
Will give it another try then with a later version or compile myself.
I’ll shout if something still does not work

[marcel_verpaalen (Marcel)]

Indeed Chris,
Very well possible that this was the issue

Cloudbees has stopped building

Looks like there’s an error on the server: https://openhab.ci.cloudbees.com/job/openHAB/scmPollLog/ Started on Aug 25, 2015 3:01:00 AM Using strategy: Default [poll] Last Built Revision: Revision 1fb0c00f548e2581ff32e879d284eb349ad5b226 (refs/remotes/origin/master) ERROR: Failed to record SCM polling for hudson.maven.MavenModuleSet@13413aa9[openHAB] java.lang.NullPointerException [...snip...] The last successful build was on 13th August.

[marcel_verpaalen (Marcel)]

Well, I was digging bit more into the problem and I’m bit further.
Looks after all that this was not a compile issue. Even with recompiling it did not work.
However if I used the new 1.8 jar in my production setup (still on pre- R1.7) it just did work well.

I traced the crash to the very first line when the serial lib is invoked here
So far it looks to me that my board has bit of issue with the serial lib as that seems the only deviation.
in my working environment I’m still on lib/nrjavaserial-3.8.8.jar however both 1.8 and OH2 are using lib/nrjavaserial-3.9.3.jar

Next step is to put an old org.openhab.io.transport.serial back to see if that indeed fixes my issue