pagenodes deployment failed.
ZoomyCat opened this issue · comments
I was trying to make a docker file for pagenodes and the installation failed. I cannot seem to get this to deploy with the given instructions.
I have attached the commands and debug information below.
made a lot of changes recently, could you give it another shot?
Will check it out again and report back.
I tried again and docker checks the return status of the things that happen. Since there are tests done to determine which OS things need when the OSX darwin tests are done this causes the docker build to fail.
Here is the output for the install stage from a manual entry into a docker container.
debug.txt
Basically what happens in docker is the build stage will see anything other than a 0 exit code as a failure to complete that stage. Since you produce a 1 exit status on Linux not be OSX its always going to fail to install inside an automated docker build afaik.
The different contexts of docker make it difficult to give exact debug information because the actual npm debug.log is stored inside the docker container which goes bye bye the second the build ends and since we have a exit status of 1 it WON'T move onto the next stage so you can't even use cat to force the debug output.
Dockerfile.txt
Dockerfile1.txt
Here are the two dockerfiles I've used in attempting this process, Github won't let me just upload them without renaming so to test each one rename them to "Dockerfile".
I put this file in the main pagenodes root and then run docker build as described below to do the build.
docker build --no-cache --rm=true -t pagenodes .
The cmd line of the Dockerfile may not be correct, It may need to be changed to have an entrypoint passing the arguments or something but without a passing install I could not test further.
Dockerfile should do a real build, Dockerfile1 creates a testing environment to run the install commands inside docker. This does not automated properly and is only a debugging stage.
Docker building like this helps with deployment, If you have a deployable method via docker then it does not matter what OS someone uses then your software is cross-platform. People just need to have docker.