On premise pipeline - JavaScript heap issues


I am currently evaluating both the point-cloud-tiler and the model-tiler tools from the in house pipeline (v4.4.2). The point-cloud-tiler seems to be able to utilise more than 2GB of RAM per core, however the model-tiler will not. When trying to process large OBJ files the process crashes with the error “CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory”. I’ve tried to resolve this by setting the max-old-space-size node option, but this causes the model-tiler to fail to run. Having set this node option to any size, you cannot even run ‘model-tiler --version’ anymore. The point-cloud-tiler however, seems unaffected. I’ve tried this now on both Windows and Linux and they both behave the same.

Any ideas on how to fix this?

Hi there,

Sorry for the delay. How large is your obj file in megabytes / gigabytes? Is it possible for you to share that obj file so we can reproduce this crash? I’ve filed an internal bug report to track this crash. Thanks.

Hi Sam,

Thank you for your reply.

The .obj file itself is 4.8GB, the total size of the model including the .png textures is 7.8GB. This model is part of a private commercial project, so I don’t want to share it publicly, but I can share it will Cesium support. Could you please contact me directly so we can find the best way to do this?

Of course. You can upload it to ion and then send us a private support message with the asset ID listed in the message. Alternatively you can include a download link to the asset from another location in the message, it’s up to you. Thanks.