jupyter lab build consistently failing due to "FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory" error

[gnestor]

[LabBuildApp] > node /Users/grant/Sites/jupyter/jupyterlab/jupyterlab/staging/yarn.js run build:prod
yarn run v1.5.1
$ webpack --config webpack.prod.config.js

<--- Last few GCs --->

[88208:0x103001a00]   122153 ms: Mark-sweep 1403.3 (1472.8) -> 1403.3 (1445.8) MB, 1875.5 / 0.0 ms  (+ 0.0 ms in 0 steps since start of marking, biggest step 0.0 ms, walltime since start of marking 1876 ms) last resort GC in old space requested
[88208:0x103001a00]   123778 ms: Mark-sweep 1403.3 (1445.8) -> 1403.3 (1445.8) MB, 1624.3 / 0.0 ms  last resort GC in old space requested


<--- JS stacktrace --->

==== JS stack trace =========================================

Security context: 0x3298f8525ec1 <JSObject>
    1: SourceMapGenerator_addMapping [/Users/grant/anaconda/share/jupyter/lab/staging/node_modules/uglify-es/node_modules/source-map/lib/source-map-generator.js:~103] [pc=0x28683eb7daab](this=0x3298b8ab5721 <SourceMapGenerator map = 0x3298474bf191>,aArgs=0x3298dd4a7de9 <Object map = 0x32984749f581>)
    2: add [0x3298db682311 <undefined>:~13790] [pc=0x28683eb84e51](this=0x32986a416c19 <Object ma...

FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory
 1: node::Abort() [/usr/local/bin/iojs]
 2: node::FatalException(v8::Isolate*, v8::Local<v8::Value>, v8::Local<v8::Message>) [/usr/local/bin/iojs]
 3: v8::internal::V8::FatalProcessOutOfMemory(char const*, bool) [/usr/local/bin/iojs]
 4: v8::internal::Factory::NewFillerObject(int, bool, v8::internal::AllocationSpace) [/usr/local/bin/iojs]
 5: v8::internal::Runtime_AllocateInTargetSpace(int, v8::internal::Object**, v8::internal::Isolate*) [/usr/local/bin/iojs]
 6: 0x28683e90463d
 7: 0x28683eb7daab
 8: 0x28683eb84e51
 9: 0x28683eb87f2e

[gnestor]

I've observed that this can be avoided by using the --dev flag when running jupyter lab build.

[jasongrout]

You can also increase the memory allocation for node by setting this environment variable:

export NODE_OPTIONS=--max-old-space-size=4096

[jasongrout]

(This happened to me when I had a huge js file in a plugin. Separating out the js file helped.)

[gnestor]

Thanks @jasongrout!

[gnestor]

I have installed MANY labextensions and I may have hit the limit of what our webpack build can handle. Using NODE_OPTIONS=--max-old-space-size=8192 jupyter lab build (from what I can tell, 8192 is the max) consistently fails with a CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory error. I am able to workaround this with NODE_OPTIONS=--max-old-space-size=8192 jupyter lab --watch.

Here's my list of installed labextensions:

$ jupyter labextension list
JupyterLab v0.32.0
Known labextensions:
   app dir: /Users/grant/anaconda/share/jupyter/lab
@hadim/jupyterlab_gist
        @hadim/jupyterlab_gist v0.3.1  enabled  OK*
@jupyter-widgets/jupyterlab-manager
        @jupyter-widgets/jupyterlab-manager v0.35.0  enabled  OK
@jupyterlab/facets-extension
        @jupyterlab/facets-extension v0.16.0 disabled OK
@jupyterlab/fasta-extension
        @jupyterlab/fasta-extension v0.16.0  enabled  OK*
@jupyterlab/geojson-extension
        @jupyterlab/geojson-extension v0.16.0  enabled  OK*
@jupyterlab/github
        @jupyterlab/github v0.6.0  enabled  OK
@jupyterlab/google-drive
        @jupyterlab/google-drive v0.12.0  enabled  OK
@jupyterlab/katex-extension
        @jupyterlab/katex-extension v0.16.0  enabled  OK*
@jupyterlab/plotly-extension
        @jupyterlab/plotly-extension v0.16.0  enabled  OK*
@jupyterlab/vdom-experimental-extension
        @jupyterlab/vdom-experimental-extension v0.15.7  enabled  OK*
@jupyterlab/vega2-extension
        @jupyterlab/vega2-extension v0.16.0  enabled  OK*
jupyterlab-chart-editor
        jupyterlab-chart-editor v0.2.0  enabled  OK*
jupyterlab-kernelspy
        jupyterlab-kernelspy v0.1.3  enabled  OK
jupyterlab-monaco
        jupyterlab-monaco v0.1.0  enabled  OK*
jupyterlab-toc
        jupyterlab-toc v0.2.1  enabled  OK
jupyterlab_voyager
        jupyterlab_voyager v0.1.0  enabled  OK*
nbdime-jupyterlab
        nbdime-jupyterlab v0.2.0  enabled  OK
qgrid
        qgrid v1.0.1  enabled  OK

[jasongrout]

Can you try --trace-gc-verbose (from https://stackoverflow.com/questions/16586705/configuring-v8s-memory-management-to-be-smart-for-a-node-js-process) to see if we can figure out what the memory situation looks like?

[ThatAIGeek]

So what's the state of this problem? Just experienced it with 0.33.11

[blink1073]

0.34 uses dev_mode by default. We have an rc out now, final release is today.

[saulshanabrook]

Closing this since we use dev_mode by default now in 0.34 and that should hopefully fix this