Hello,
First thanks to the team for developping such a convenient shiny app deployment solution.
I have an app in production with kubernetes on GCP and using v2.6.1. The typical start time of the app after login is 10 seconds.
I try to switch to v3.0.1 with the same configuration, including the application.yml and only changing the shinyproxy jar version.
The app is not starting and I get the usual message “Failed to start container”.
So I added container-wait-time: 90000 to the yml.
Now the app is starting but it takes nearly 3 minutes !
The shinyproxy container log says “Container unresponsive, trying again (37/55)”, incrementing every 7 seconds, and finally I get the “Proxy activated” log message.
Would you have an idea why it takes so long with the new v3.0.1 version ? Is there a parameter wrongly set up in the yml for instance ?
Thanks in advance for your time and advice !
Hi @Sylvain_Guilbert,
I never had problem like this before, although I was busy with this “Failed to start the container” error for four days! I think I solved my problem, I hope it help you too.
when I was running docker run -p 3838:3838 image_name apparently there was no problem, also I couldn’t access the application on the browser on any port. when I was running that I was getting this: listen 127.0.0.1:someportnumber
it supposed to be 0.0.0.0:someportnumber though! The solution is the way you call the application in the application.yml to force it go to host 0.0.0.0 and port 3838. I added the host and port and it worked. something like this:
container-cmd: [“R”,"-e",“shiny::runApp(’~/yourappfolder’,host=‘0.0.0.0’,port=3838)”]
The rest of the application.yml doesn’t need a change and I don’t think we need any other setting to make it work.
take care,
Mehrdad
Thanks @Mehrdad_Varedi
My app is a package developed with Golem.
I start the app with
container-cmd: ["R", "-e", "options('shiny.port'=3838,shiny.host='0.0.0.0');myapp::run_app()"]
In the pod log I indeed have Listening on http://0.0.0.0:3838
This works fine in v2.6.1 but takes very long in v3.0.1, and again the only difference between my two tests is the jar file.
Hi Sylvain,
It seems that we are facing a similar issue. ShinyProxy makes several attempts every 7 seconds, quoting “Container unresponsive, trying again (37/55)”, while we notice that the containers are running successfully in the background - but cannot get connected/seen by shinyproxy.
Have you managed to find a working solution to this?
Thanks in advance.
Hi @andrewp , this usually means the network connection between ShinyProxy and the containers it not properly working.
Can you open a new topic and include which container backend you are using and your shinyproxy config?