Docker Community Forums

Share and learn in the Docker community.

Docker won't start after Mac OS X update - 10.11.5


(Ramon Bosch Smit) #1

I just got the update for OS X 10.11.5. Now docker won’t start. I have restarted the OS a couple of times and the outcome is the same. Docker constantly reports to be 'starting.

syslog -k Sender Docker produces nothing; ending in a ‘docker is not responding’.

The docker for Mac beta version I’m running is 1.11.1 - beta 13.1. This was working before the update. Nothing else appears to have changed.


(Anil Madhavapeddy) #2

Hi Ramon,

Unable to repro this one – I just updated a Mac from 10.11.4 to 10.11.5 with beta 13.1 and it came up fine. Would you be able to run pinata diagnose -u from a terminal and send us the UUID to beta-feedback@docker.com? We can try to trace the issue down from them hopefully.

You might also want to try a “reset to factory defaults” after that, which will trigger a fresh reinstall the next time you run it. That’ll hopefully flush out whatever is blocking the application start. We’d appreciate getting a copy of the broken logs before you do this, so that we can try and track down the issue.


(Mzollneritsch) #3

My docker also stucked after update. A reset helped but I lost all my data container.

Here is my pinata output:

    [maxzollneritsch:~] 9m22s $ pinata diagnose -u
    OS X: version 10.11.5 (build: 15F34)
    Docker.app: version: v1.11.1-beta14.4.m
    Running diagnostic tests:
    Error docker ps: timeout after 10.00s
    [ERROR]   docker-cli       docker ps failed: Failure("docker ps: timeout after 10.00s")
    [OK]      docker-cli
    [ERROR]   Moby booted      /Users/maxzollneritsch/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/console-ring does not exist
    [OK]      Moby booted
    [OK]      driver.amd64-linux
    [OK]      vmnetd
    [OK]      osxfs
    [OK]      db
    [OK]      slirp
    [OK]      menubar
    [OK]      environment
    [OK]      Docker
    [OK]      VT-x
    Error docker ps: got timeout (docker ps: exit 1)
    Error exec: echo "00000003.0000f3a6" | nc -U /Users/maxzollneritsch/Library/Containers/com.docker.docker/Data/@connect > /tmp/20160606-154810/diagnostics.tar: exit 1
    Docker logs are being collected into /tmp/20160606-154810.tar.gz
    Most specific failure is: /Users/maxzollneritsch/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/console-ring does not exist
    Your unique id is: E94B1590-7D64-4B71-8E7D-1677C9DA6196
    Please quote this in all correspondence.

(Ramon Bosch Smit) #4

Hi Anil,

thank you for the prompt response. Regrettably I did reset Docker to factory settings right away, without running pinata first.

The issue I have now is that I can’t use Docker anymore, as I can’t seem to find the beta token. That has obviously nothing to do with the tech or this forum :slight_smile:


(Guilhermefalcao) #5

Same problem here. Is the first time that i have installed Docker on my Mac. But after install finish, docker dont start.
Any solution?


(Karllhughes) #6

Same problem here with similar output from pinata as @mzollneritsch posted.

I filed a ticket with docker support, but I guess until there’s a fix I’m sort of stuck.


(Joseph Crotty) #7

+1 to a fix as my pinata output is similar to others. This is a fresh install on a mac book pro that is ~3months old. Previously to installing Docker Beta I had docker, docker-machine and docker-compose all installed w/ MacPorts. I removed all the ports before attempting to install Docker Beta, but did not reboot. After I installed Docker Beta I get the pinata output below. However, after I restarted my computer Docker Beta started working 100% so far and pinata cleared up to all green success codes. I don’t get why a reboot was necessary. Also, I could not install Docker Beta as an unprivileged user which was a PITA as it’s a corp macbook pro and I had to jump through IT hoops to get it installed.

$ pinata diagnose -u
OS X: version 10.11.5 (build: 15F34)
Docker.app: version: v1.11.1-beta14.5.m
Running diagnostic tests:
Error docker ps: timeout after 10.00s
Error docker ps: timeout after 10.00s
[ERROR] docker-cli docker ps failed: Failure(“docker ps: timeout after 10.00s”)
[OK] docker-cli
[ERROR] Moby booted /Users/jcrotty/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/console-ring does not exist
[OK] Moby booted
[OK] driver.amd64-linux
[OK] vmnetd
[ERROR] osxfs com.docker.osxfs is not running
[OK] osxfs
[OK] db
[OK] slirp
[OK] menubar
[OK] environment
[OK] Docker
[OK] VT-x
Error exec: echo “00000003.0000f3a6” | nc -U /Users/jcrotty/Library/Containers/com.docker.docker/Data/@connect > /tmp/20160608-164655/diagnostics.tar: exit 1
Docker logs are being collected into /tmp/20160608-164655.tar.gz
Most specific failure is: com.docker.osxfs is not running
Your unique id is: E7E384C5-BAD9-400E-9204-4101F59C01CD
Please quote this in all correspondence.


(Rohinton Kazak) #8

Also you were not running docker toolbox and friends previously?


(mdelataille) #9

I have same problem. Did you find a solution ?


(Lucas Chaufournier) #10

I am getting similar timeout issues.
I used the following command: "screen ~/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/tty"
to enter the docker virtual machine on OSX as root. When I ran service docker status it says: * status: crashed. It seems this is why it is unable to connect and we are getting timeouts. Unfortunately, I still haven’t discovered the cause of why it is crashing. Maybe this will help someone further debug the issue.

Edit: Also when the VM boots up it appears to display the following:
✓ Drive found: vda2016/06/09 15:42:31 Failed to bind to hvsock port: 0x61

✓ Drive mounted: /dev/vda2 on /var type ext4 (rw,relatime,data=ordered)
✓ Network connected: inet addr:192.168.65.2 Bcast:0.0.0.0 Mask:255.255.255.248
✓ Process transfused running
✗ No tap-vsockd process
✗ No docker process
✓ Process containerd running: docker-containerd -l /var/run/docker/libcontainerd/docker-containerd.sock --runtime docker-runc --debug --metrics-interval=0
✗ Docker ps failed: Cannot connect to the Docker daemon. Is the docker daemon running on this host?
✓ Diagnostics server running: /usr/bin/diagnostics-server

  • Adjusting oom killer settings … [ ok ]

(Anil Madhavapeddy) #11

We think that one root cause of this is due to an unclean shutdown of the Linux VM leading to filesystem corruption of the AUFS layer. We should never have an unclean shutdown in the normal course of events (autoupdates, application shutdown, etc…) so we are still diagnosing why this happens.

You can verify if it is AUFS corruption by checking the docker.log on your Mac by running:

tail ~/Library/Containers/com.docker.docker/Data/com.docker.driver.amd64-linux/log/docker.log

If you see ominous AUFS-related messages, then your best bet is to reset to factory defaults (by going to the Preferences menu), while will delete the container storage and reinitialise everything from scratch. Be sure that you didn’t have any container images that you don’t mind losing before you do this.


(Karllhughes) #12

The latest update (1.11.2-beta15) fixed the issue for me.


(Hieuletrung) #13

Mine was corrupted
time="2016-09-18T04:24:29.038861061Z" level=fatal msg="Error starting daemon: error initializing graphdriver: mkdir /var/lib/docker/aufs: not a directory"

After reset factory, it can start now.


(Aviral75) #14

Yes Reset to factory default , it then found i was using old images and migrated them and thereon things look find like docker ps etc respond back.