This is basically a place to try and document what I have done, and find out what I need do in order to get this working. I've flagged it as pine64 and pinebook related as ayufan's builds support both at this point in time.
So, what am I trying to do? Basically, I have crappy internet, and come computers around which are more than capable of grunting their way through compiling and building images for the pine64 / pinebook. This would also let me look into customising the images, and doing some work on them myself... maybe.
I've played with docker a little bit, and have used Travis CI in the past, so thought how hard could this be? Famous last words. 
I decided one of my pine64 servers could be the jenkins boss, so loaded jenkins 2.46 (as it's the LTS version) on that that. Easy peasy, had everything up and running in under 20 minutes. Set up a build node or two on the ubuntu x64 boxes, so that they could be responsible for the actual builds, so the pine64 only had to manage them from afar. Since the Jenkins file in the repo wanted a node with both the docker and linux-build tags, I made sure one of them had that, was running LTS 16.04, and had docker on it, and had actually also manually build the build-environment docker on it also (but that was more to test if it was working than actually configure anything, and I did end up deleting it at one point just in cause it was the cause of the issue).
I then setup the new pipeline, configured it as a pipeline script from SCM -> Git -> https://github.com/ayufan-pine64/linux-build and left pretty much everything else as defaults. Tried running the pipeline, had to double back and install the AnsiColor plugin, and then the build ran. It downloaded the repo file, logged into the build node, and started building the docker container without any dramas. So it gets to the end of the docker build, and promptly and consistently dies. I'll include some logs below, but does anyone (especially ayufan or tkaiser) have any ideas what I'm doing wrong? Am I missing some blindingly obvious configuration step that I should have known about before attempting any of this?
I even setup jenkins on the same machine that was doing the build, and set it as the docker/build node, tried playing with docker plugins with no luck, and installed blue ocean just incase it was related to a plugin it added.
I *think* the issue is related to this entry in the log, but I'm not sure, and I don't know how to correct it, hence this post. asus-h55d is the host-name of the node which the build was slaved out to, and it's this line that stands out as different from the logs from ayufan's builds.
This is the plain text build log: http://sprunge.us/LaPW
This is the log that blue ocean shows, which rather helpfully doesn't actually show why the environment stage failed: http://sprunge.us/IiVU
Can anyone help?
So, what am I trying to do? Basically, I have crappy internet, and come computers around which are more than capable of grunting their way through compiling and building images for the pine64 / pinebook. This would also let me look into customising the images, and doing some work on them myself... maybe.
I've played with docker a little bit, and have used Travis CI in the past, so thought how hard could this be? Famous last words. I decided one of my pine64 servers could be the jenkins boss, so loaded jenkins 2.46 (as it's the LTS version) on that that. Easy peasy, had everything up and running in under 20 minutes. Set up a build node or two on the ubuntu x64 boxes, so that they could be responsible for the actual builds, so the pine64 only had to manage them from afar. Since the Jenkins file in the repo wanted a node with both the docker and linux-build tags, I made sure one of them had that, was running LTS 16.04, and had docker on it, and had actually also manually build the build-environment docker on it also (but that was more to test if it was working than actually configure anything, and I did end up deleting it at one point just in cause it was the cause of the issue).
I then setup the new pipeline, configured it as a pipeline script from SCM -> Git -> https://github.com/ayufan-pine64/linux-build and left pretty much everything else as defaults. Tried running the pipeline, had to double back and install the AnsiColor plugin, and then the build ran. It downloaded the repo file, logged into the build node, and started building the docker container without any dramas. So it gets to the end of the docker build, and promptly and consistently dies. I'll include some logs below, but does anyone (especially ayufan or tkaiser) have any ideas what I'm doing wrong? Am I missing some blindingly obvious configuration step that I should have known about before attempting any of this?
I even setup jenkins on the same machine that was doing the build, and set it as the docker/build node, tried playing with docker plugins with no luck, and installed blue ocean just incase it was related to a plugin it added.
I *think* the issue is related to this entry in the log, but I'm not sure, and I don't know how to correct it, hence this post. asus-h55d is the host-name of the node which the build was slaved out to, and it's this line that stands out as different from the logs from ayufan's builds.
Code:
[Pipeline] withDockerContainer
15:33:39 asus-h55d does not seem to be running inside a containerThis is the plain text build log: http://sprunge.us/LaPW
This is the log that blue ocean shows, which rather helpfully doesn't actually show why the environment stage failed: http://sprunge.us/IiVU
Can anyone help?
