0

My build fails without any useful information. The problem started when I added ethers library as a dependency to my package.json file.

Here's output from web console:

Pulling image "registry.access.redhat.com/rhscl/nodejs-6-rhel7@sha256:0860a4ccdc062f5ab05ec872298557f02f79c94b75820ded9a16211d8ab390ce" ...
---> Installing application source ...
---> Building your Node application from source
> grpc@1.7.1 install /opt/app-root/src/node_modules/grpc
> node-pre-gyp install --fallback-to-build --library=static_library

Command oc status -v gives nothing usefull either.

All I can say is that the issue has something to do with node-pre-gyp or node-gyp compiling. I'm totally out of my depth here. Any advice?

manidos
  • 3,244
  • 4
  • 29
  • 65

1 Answers1

1

The npm installer can at times chew up quite a lot of memory when doing installs and require more than the default memory resources given to a build pod. When it fails due to that, there aren't really any messages in the logs. You may see something in events under monitoring.

The default memory resources for a build is 512Mi. See following answer which deals with similar issue for Python:

for description of how to increase memory available to a build.

Graham Dumpleton
  • 57,726
  • 6
  • 119
  • 134
  • @GrahamDumpletom, You're absolutely right. Here are event logs: `Error creating: pods "olmeo-13-" is forbidden: exceeded quota: compute-resources, requested: limits.cpu=1,limits.memory=512Mi, used: limits.cpu=2,limits.memory=1Gi, limited: limits.cpu=2,limits.memory=1Gi`. Thanks! – manidos Nov 15 '17 at 05:35
  • That error you are getting now indicates that limit request was larger than quota. What OpenShift cluster is this on? Even Online Starter should have 1Gi for memory maximum for terminating resources such as builds. So that should work and you shouldn't get that error. If you had other terminating resources claimed, it should have just waited until they were released. – Graham Dumpleton Nov 15 '17 at 06:10