openHAB 1.8 and 2.0 beta1 Release is out!

My HABmin stopped working too. I just reinstalled it and then it worked fine.

1 Like

Thanks @theo,

I will also change it in the Ubuntu install guide on the wiki.

3 Likes

Iā€™m currently running openhab 1.8.0 installed from apt-repository on my raspberry .
Can I install an openhab 2.0 instance on the same raspberry ?

why not (assuming your are using a rip 2) ā€¦ it just depends on your patience when starting both instances at the same time. Please note you will have to change ports of one of the installations (see start.sh for details).

But if you have devices accessed by any OH instances (for example, zwave), only the first one to start will get access to it and not the second one.

HTH

Thanks a lot @theoā€¦ this is awesome! It also provides a good place to specify DEBUG option/mode.

Since upgrading from the 1.8 Nightly dated 10/8/15 to the released 1.8 My log has
[WARN ] [g.openhab.io.net.exec.ExecUtil] - Process exited with an error: 0 (Exit value: 0)
added every time I run a command line Curl from a rule.
The rule still works so Iā€™ve added

<logger name="org.openhab.io.net.exec.ExecUtil" level="OFF" />

to logback.xml
Iā€™m guessing this is a bug you need to know about?
The line from the rule that seems to cause the error is

 var BedroomHeatpumpStatus = executeCommandLine('curl -sSH "Accept: application/json" "https://home.sensibo.com/api/v2/pods/MyPodID/acStates?apiKey=MyApi&limit=1&fields=acState"', 5000)

I suspect that these unwanted log entries are the result of this commit.

@kevin, could you open a new issue in github?

@steve1, could you take a look at undoing these unwanted log entries for 1.8.1?

Interesting that they are logged as ā€œWARNā€ rather than "DEBUG."
Iā€™ll log an issue when Iā€™m near a desktop tomorrow

Thanks Kevin.

@kevin and @watou. That message is only logged if there was an exception resulting from the execution. Do we know why thereā€™s an exception in this case?

I understand that was the intention of your PR, but it appears that running curl is returning with exit code 0 (normal), and @kevin was not reporting that his curl command was failing, so it appears that the logging is happening regardless. Do I have that right, @kevin?

[WARN ] [g.openhab.io.net.exec.ExecUtil] - Process exited with an error: 0 (Exit value: 0)

I also am of the opinion that users may intentionally expect non-zero return codes, so in the case where those would log WARN messages, users would see that as unhelpful log pollution. I suggest either reverting that referenced commit, or changing logging to DEBUG so at least users can by default not have these messages logged.

This is very similar to HttpUtil putting WARN messages if a non-200 HTTP status code comes back. Itā€™s also incorrect to log these at WARN level, because many callers would entirely reasonably expect non-200 return values. (I changed this in ESH.)

Created Issue #3821 for further discussion.

@watou. I think we should understand the cause of the problem better before making a decision. I also believe that masking exec exceptions will cause the user more problems than warning about them. I see many posts in this forum from people with exec binding issues. Many of them do not run with debug logging turned on and some donā€™t even know how to configure the logging to get that output.

The Apache Commons Executor does not, by default, raise exceptions on any exit code. This behavior is being configured by openHABā€™s ExecUtil for some reason. Iā€™m going to investigate that further and do some experimentation. Once we have that information, we can make an informed decision about how to respond.

1 Like

No errors as far as I can see, that rule has been running for weeks now and nothing in the log until the change to V1.8 final.
Any errors cause a stack dump with the next line in the rule which I change into something sensible with ā€œCatch Throwable Tā€

@Steve1, the rule has been running happily for weeks now and never put anything into the log until I upgraded from V1.8 Nightly to V1.8 released. Iā€™ll pop over to the issue page to continue the discussion.

Thanks for the update. It works fine, except that the TCP binding throws an errors, when using the ā€˜MAP (filename)ā€™ command, it cannot find the file, although the file is there. I downgraded the TCP binding to 1.7, and everything is fine now. But a new error seem to have introduced is this area in the TCP binding in 1.8.

Regards Lars Oksbjerre

@loksbjerre, do you have a related log file entry?

Hi, here it is:

2016-01-20 12:32:13.054 [ERROR] [t.i.s.MapTransformationService] - opening file 'M1.MAP)'], >[0:192.168.2.102:12345:'MAP(M1.MAP' throws exception
java.io.FileNotFoundException: configurations/transform/M1.MAP)'], >[0:192.168.2.102:12345:'MAP(M1.MAP (No such file or directory)
	at java.io.FileInputStream.open(Native Method) ~[na:1.7.0_65]
	at java.io.FileInputStream.<init>(FileInputStream.java:146) ~[na:1.7.0_65]
	at java.io.FileInputStream.<init>(FileInputStream.java:101) ~[na:1.7.0_65]
	at java.io.FileReader.<init>(FileReader.java:58) ~[na:1.7.0_65]
	at org.openhab.core.transform.internal.service.MapTransformationService.transform(MapTransformationService.java:60) ~[na:na]
	at org.openhab.binding.tcp.protocol.internal.UDPBinding.transformResponse(UDPBinding.java:261) [bundlefile:na]

@loksbjerre, the first 10 lines of that output is all I need. Will you edit the post to delete the rest? Otherwise, itā€™s more difficult for people to follow the discussion (lots of scrolling).

Also, can you post the full item definition (or at least the binding configuration part) so I can test it once I have access to my dev environment. Thanks.

Could you guys please use the ā€œReply as linked Topicā€, so that this topic is not flooded with off-topic stuff? Thanks!