Com.docker.xhyve and com.docker.osxfs cpu usage

Same issue here. I am using OSX 10.11.4 (15E65) . I am using a compose file with 3 Meteor apps , 1 Node.js and 1 python containers. Getting 250% to 300% of CPU usage on com.docker.hyperkit

❯ pinata diagnose -u OS X: version 10.11.4 (build: 15E65) Docker.app: version v1.11.1-beta11 Running diagnostic tests: [OK] docker-cli [OK] Moby booted [OK] driver.amd64-linux [OK] vmnetd [OK] osxfs [OK] db [OK] slirp [OK] menubar [OK] environment [OK] Docker [OK] VT-x Docker logs are being collected into /tmp/20160513-132631.tar.gz Most specific failure is: No error was detected Your unique id is: B123951D-7DFD-49DD-87D9-73832E062133

Same here, with an idling Django app. Auto-reload is on (vanilla runserver command), but no files are changing and I’m not viewing the site. com.docker.hyperkit bounces between 70% and 135%, and com.docker.osxfs from 25-40%. Stopping that container (and leaving my other docker-compose containers running) makes them both drop to under 1%.

Edit: I followed the advice here High CPU usage for Django application and installed pyinotify. Fixed it!


When running a watch gulp task on a shared volume between host and guest. I can give samples of my environment if needed.

The same for my docker-compose project of 7 containers (RoR app) from cold MBP start (not after waking up from sleep mode).

Diagnose when running the beta:

OS X: version 10.11.5 (build: 15F34)
Docker.app: version v1.11.1-beta13.1
Running diagnostic tests:
[OK]      Moby booted
[OK]      driver.amd64-linux
[OK]      vmnetd
[OK]      osxfs
[OK]      db
[OK]      slirp
[OK]      menubar
[OK]      environment
[OK]      Docker
[OK]      VT-x
Docker logs are being collected into /tmp/20160605-145926.tar.gz
Most specific failure is: No error was detected
Your unique id is: 1B8DEAC3-B865-45CC-BACE-E1A7A6A409F9
Please quote this in all correspondence.

The system: MBP mid 2014, Core i7 2.2 Ghz, 16 Gb RAM DDR3, OS X El Capitan 10.11.5.

The same project running on docker 1.11.1 within VirtualBox consumes only 7% system and 2.63% user CPU load (vs 39.10% & 13.38% correspondingly from the screenshot).

Any updates from engineering on this? Is any additional community input needed for reproducing and analyzing this? I’m experiencing this several times per day and it’s basically a deal breaker for the beta. I may have to go back to docker toolbox until this issue is addressed.

I am having the same issue as of the latest update. From a fresh boot, after I do a docker-compose I am at over 300% CPU usage. Here is my diagnostic ID: 44B9FF1E-0592-44E1-BA0E-34687F195BC2

It’s not ideal but that would be a whole lot better than my current situation. I tried searching for this --noreload option but couldn’t figure out which program it applies to. Can you provide a bit more detail?

After pulling my first docker image and running. com.docker.xhyve and com.docker.driver.amd64-linux were high until I killed the container. The container was using a volume ‘run -v’. After removing the bind mount volume the problem was fixed. HTH for other that encounter the same problem.

@bottybotbot can you describe the command to remove the bind mount volume? I’d like to try that but I’m a docker noob.

@jon47 I was mapping a local volume for the app to store data.

for example: ‘-v /my/local/data:/my/docker/data’

Same problem with with Docker for Mac 1.12.0-rc4-beta20 (build: 10404).

I have a shared volume defined in docker-compose for live reloading code for development. When starting the container com.docker.hyperkit is running at 259% and com.docker.osxfx is running at 140% +.

My containers are a MeanJs code base and Mongo container. The shared volume in docker-compose is:

volumes:

  • .:/home/mean:rw