我如何在 ParcelJS 中找到 memory 泄漏（或什么占用了这么多内存）？

Question

I'm running ParcelJS 2 in a BitBucket pipeline bundling up an app, but our pipelines sporadically runs out of memory.

I want to be able to pin-point where the problem comes from - whether that is a single node module dependency, or if it's just a problem with ParcelJS overall.

I've tried logging out the memory usage in the CI container while the build is taking place, and I can see it's hitting the max memory towards the end of the build - sometimes this causes a crash and sometimes doesn't. I've done that by running this command in the pipeline:

- while true; do ps -aux && sleep 30; done &
- while true; do echo "Memory usage in megabytes:" && echo $((`cat /sys/fs/cgroup/memory/memory.memsw.usage_in_bytes | awk '{print $1}'`/1048576)) && sleep 0.1; done &

Which outputs...

4082
Memory usage in megabytes:
4083
Memory usage in megabytes:
4093
Memory usage in megabytes:
4095
Memory usage in megabytes:
4095
Memory usage in megabytes:
4094
Memory usage in megabytes:
4095

etc...

Any advice on how to go about pin-pointing where the problem could be?

Thanks in advance.

Answer 1

This is a good command to use and it seems to work fine for Bitbucket Pipelines and possibly other CI/CD services, like GitHub Actions and GitLab CI/CD. It displays the memory usage inside the Docker container, in real-time, during a CI/CD build.

Here are some references and tutorials about the Docker memory usage commands:

• https://confluence.atlassian.com/bbkb/troubleshooting-bitbucket-pipelines-1141505226.html
• https://fabianlee.org/2020/01/18/docker-placing-limits-on-container-memory-using-cgroups/
• https://github.com/oracle/docker-images/issues/1939

As you already know, Bitbucket Pipelines has a strict limit on the memory usage of each Docker container. By default, each build step has 4 GB of memory available:

• https://support.atlassian.com/bitbucket-cloud/docs/configure-bitbucket-pipelinesyml/
• https://janosmiko.com/blog/2021-09-18-demystifying-bitbucket-pipelines-memory-limits/

To pin-point the large memory use, consider changing the sleep time to 2 or 5 seconds, and perhaps add a timestamp to each message, using the " date " command. Then check which command was run at the time of the large memory increase. Double-check if this command is creating too many threads or processes, or if it's opening too many files at the same time. The command may have some flags that you can use to reduce the number of threads or processes that it creates.