Docker Community Forums

Share and learn in the Docker community.

Beta9 broke Jekyll :(


(Morendil) #1

Expected behavior

Configuration file: /srv/jekyll/_config.yml
            Source: /srv/jekyll
       Destination: /srv/jekyll/_site
 Incremental build: disabled. Enable with --incremental
      Generating... 
                    done in 1.885 seconds.
 Auto-regeneration: disabled. Use --watch to enable.
ok: down: /etc/startup3.d/nginx: 0s, normally up

(Observed behavior in Docker Toolbox + VirtualBox, was working in beta8)

Actual behavior

           Source: /srv/jekyll
      Destination: /srv/jekyll/_site
Incremental build: disabled. Enable with --incremental
     Generating... 

…and the process hangs there, apparently indefinitely

Information

OS X: version 10.11.1 (build: 15B42)
Docker.app: version v1.11.0-beta9
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/20160428-232014.tar.gz
Most specific failure is: No error was detected
Your unique id is: 3A90ADC3-0486-4B11-9BC9-51D7B3BE367C
Please quote this in all correspondence.

Steps to reproduce the behavior

cd into an empty directory, then

docker run --rm --label=jekyll -v $(pwd):/srv/jekyll -it -p 4000:4000 jekyll/jekyll:3.0.3 jekyll build

(David Sheets) #2

I cannot reproduce this in an empty directory but I can reproduce this behavior in my home directory. Does this happen for you in an empty directory? When the symptom occurs, do you experience high CPU load? What kind of configuration related to the file system is present in _config.yml?

Jekyll appears to be walking the entire subtree at which it’s pointed. We are currently working on improving directory traversal performance and shared volume latency so that this use case is not intolerably slow. Do you have a project directory that is large enough to experience this slowdown?


(Morendil) #3

I’m on a different machine now, so I can’t try repro’ing under the same conditions. I bravely installed beta9 on this box too, and

  • it works in an empty directory

  • it works in the local copy of the same repo it failed on yesterday

    OS X: version 10.10.5 (build: 14F27)
    Docker.app: version v1.11.0-beta9
    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/20160429-123533.tar.gz
    Most specific failure is: No error was detected
    Your unique id is: 424324DB-6BED-443D-B823-3F450617902D
    Please quote this in all correspondence.

I will try repro’ing in an empty directory when I get home tonight.


(Morendil) #4

Cannot repro in either condition now (3A90ADC3-0486-4B11-9BC9-51D7B3BE367C).


(Esilverberg) #5

I use jekyll with Docker on my OSX machine, and using Docker beta 1.11.1-beta12 (build: 7528) it is brutally slow processing trivial changes:

Regenerating: 1 file(s) changed at 2016-05-24 23:10:19 ...done in 41.370099425 seconds.

Same change would take seconds previously with docker + vbox.

I see that dsheets says:

We are currently working on improving directory traversal performance and shared volume latency so that this use case is not intolerably slow

Will be very excited to get this fix!