Memory leak in com.docker.osxfs

[tverboon]

Very high memory usage:

I am running sabnzbd in a Docker container. It looks like all disk IO, that should be a stream, stays in Memory.

Docker for Mac: version: 18.03.0-ce-mac60 (dd2831d4b7421cf559a0881cc7a5fdebeb8c2b98)
macOS: version 10.13.4 (build: 17E199)

Referencing #1815. Although this is probably a new issue, previous version didn't cause excessive memory usage in my use case.

[djs55]

@tverboon thanks for your report! Could you provide reproduction steps to help me replicate this in-house? Ideally a Dockerfile or a docker-compose.yml or a script. Thanks!

[tverboon]

docker run -d --rm --name=sabnzbd \
-v /T5/nzbd/config:/config \
-v /T5/nzbd/complete:/complete \
-v /T5/nzbd/incomplete:/incomplete \
-v /T5/nzbd/watch:/watch \
-v /T5/nzbd/:/datadir \
-v /T5/nzbd/media:/media \
-p 8080:8080 \
sabnzbd/sabnzbd

But this is the easy part. To reproduce it, you need to download an NZB from a news server. Actually that's not easy to reproduce, unless you have access to a news server.

I tried to reproduce with curl in a vanilla debian container, but that didn't work. Memory grows, but shrinks also again, which looks normal buffering to me.

--
Trying to reproduce it with sabnzbd, but I see the same behaviour as with curl. Memory usage is between 700MB and 1GB, but not like yesterday where it kept growing.

I am sorry, but it seems I cannot consistently reproduce this.

[ghost]

I prepared a docker-compose with which I can reliably reproduce the issue on my system: https://github.com/zulli73/docker_issue_2839. Please check the README.md

It suspect that this issue is caused due to a kernel related resource leak, like unclosed file handles or the like as kernel_task's memory consumption grows as well (see notes in README.md).

Probably needless to mention, but: Your system will "crash", so save any data before reproducing!

[zeyangl]

having the same issue here. Is there a gonna be a fix for this?

[docker-robott]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale comment.
Stale issues will be closed after an additional 30d of inactivity.

Prevent issues from auto-closing with an /lifecycle frozen comment.

If this issue is safe to close now please do so.

Send feedback to Docker Community Slack channels #docker-for-mac or #docker-for-windows.
/lifecycle stale

[docker-robott]

Closed issues are locked after 30 days of inactivity.
This helps our team focus on active issues.

If you have found a problem that seems similar to this, please open a new issue.

Send feedback to Docker Community Slack channels #docker-for-mac or #docker-for-windows.
/lifecycle locked