[SOLVED] Log Entries & Visual Studio Code Use After 2.4 upgrade

I’ve been having three things going on after upgrading to 2.4 last Dec. I hoped they would work themselves out, but i’m still having issues and hoping someone can provide some guidance.

After upgrading:

  1. The syntax checking in Visual Studio Code no longer works. Upon starting Code and navigating to the Conf folder, an error simply comes up in Code that says “Internal Error”. It worked fine prior to the upgrade.

  2. I get these long repeating errors in the log:

2019-02-21 08:34:30.095 [WARN ] [org.eclipse.jetty.server.HttpChannel] - /rest/sitemaps/home/home
javax.servlet.ServletException: javax.servlet.ServletException: A MultiException has 1 exceptions.  They are:
1. java.lang.IllegalStateException: ServiceLocatorImpl(__HK2_Generated_5587,5588,20550606) has been shut down

(lot of stuff…)

Caused by: javax.servlet.ServletException: A MultiException has 1 exceptions.  They are:
1. java.lang.IllegalStateException: ServiceLocatorImpl(__HK2_Generated_5587,5588,20550606) has been shut down

(lot more stuff…)

Caused by: org.glassfish.hk2.api.MultiException: A MultiException has 1 exceptions.  They are:
1. java.lang.IllegalStateException: ServiceLocatorImpl(__HK2_Generated_5587,5588,20550606) has been shut down

(more stuff…)

Caused by: java.lang.IllegalStateException: ServiceLocatorImpl(__HK2_Generated_5587,5588,20550606) has been shut down

(Errors repeat)

  1. When in PaperUI or BasicUI, I will frequently get “Jersey Not Ready”, “503” or “Offline: Waiting for Connection to become available” messages.

I have two tablets that are constantly on HabPanel. Thank anyone in advance who can steer me in a direction to identify my issue.

For issue #3: Do you use a proxy? There are some challenges there. I’ve had to do some trial and error with Nginx configuration for instance. And number 2 might be related to this as well.

Thanks for your feedback. No, I am not using a proxy in my environment.

Another thing I forgot to mention is that I use one of the tablets for playing sounds via websink to HabPanel. Around when these other issues started, the tablet will now routinely miss playing a sound completely a few times a day. The rule fired play command will be present in the log, but the tablet will just miss it. I remember it being more reliable before these issues started.

To solve issue #1 you have to assigne a network drive letter.
I had the same troubles when accessing by samba share only.

Just a wild guess: do you have the REST API docs add-on installed? There have been some REST API related issue reported here that was fixed by removing the documentation add-on. :no_mouth:

Also, there are several issues on Github for Eclipse Smart Home (and OH) about Offline: Waiting for Connection to become available caused by SSE timeouts, like #597, #6403 and #5957, and not all seems to be fixed for every scenario/device/client.
What OS are the tablets running? And what browser (version)?

Sorry, I have no answer for you, only pointers and more questions…

Spot on! I added the conf folder as a mapped network drive, chose the mapped drive in Code, and boom, code validation works again! thank you so much for this nugget of knowledge!

Thanks! The REST Documentation add-on was installed. I’ve uninstalled it, and will monitor for results. One tablet is a Nexus 7 tablet, and the other is an original 7" Fire tablet. I have Silk browser open on the old Fire tablet, and Fully Kiosk running on the Nexus tablet.

In case those log entries remain, have look into This thread
I had those log entries caused by a (unwanted) legacy binding beeing installed constantly, resulting in an unresponsive system.

IMHO this warrants a reported issue on Github, or at least some documentation with the extension. Would you mind creating that issue on Github? :grin:

I’ve let this ride for almost two weeks now, and I can definitively say that this has solved my issue - I cannot thank you enough. The log entries I originally reported and web page error messages are completely gone. My system is also now running a lot smoother and more responsive. Between the guidance to remove the REST documentation and the need to add the conf directory as a network drive, the OH community has once again come through for me. Thank you all!