I have had some long existing builds in CircleCI that suddenly stopped working. These builds are for an Ionic 4/Angular 8 app. I have a small workflow, a validation stage and then android and ios build/deploy stages.
The validation stage lints, tests and runs the prod angular build. The prod angular build is what suddenly stopped working recently. The exact same commands run on my local systems just fine. They fail only when running within CircleCI. This is the blurb from the end of the log:
chunk {112} 112-es2015.48b
4e95424bf887ec877.js () 8.47 kB [rendered]
chunk {113} 113-es2015.2028f8646f217d57cfa2.js () 990 bytes [rendered]
chunk {114} 114-es2015.bc094ef077413855d22b.js () 77.1 kB [rendered]
chunk {115} 115-es2015.088b25d472e974054dc7.js () 18 kB [rendered]
chunk {116} 116-es2015.9e539c05f82d89b7863e.js () 4.11 kB [rendered]
Date: 2020-06-05T18:37:15.359Z - Hash: a773489f1c6d4dbcf844 - Time: 77844ms
Killed
npm ERR! code ELIFECYCLE
npm ERR! errno 137
npm ERR! wichopper@5.0.0-alpha.1 build:ci: `ng build --prod --no-progress`
npm ERR! Exit status 137
npm ERR!
npm ERR! Failed at the wichopper@5.0.0-alpha.1 build:ci script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.
The only information I’ve been able to find on this issue is that linux will kill a process if it needs too much memory. I’ve configured node to use the full 4096 megs of available memory on the container, as well as 2048, 1024. It does not seem to matter how much memory I allow node processes to use, the build fails the same way every time.
I’ve also done the same locally on my dev computers, increasing and reducing the amount of memory available to the process, and it always succeeds. The issue is solely extant within CircleCI. I’ve actually run the CircleCI locally with the circleci CLI, and it fails the same way when I do that as if I actually run it on CircleCI’s cloud.
So the problem seems to be directly related to CircleCI, if not directly caused by it. In any case…our build is completely dead at this point, and I’m not sure what else to do at the moment to figure out why our build process is being killed so actively and persistently by CircleCI. Any help would be greatly appreciated.