osrf / rocker Goto Github PK

View Code? Open in Web Editor NEW

530.0 22.0 68.0 259 KB

A tool to run docker containers with overlays and convenient options for things like GUIs etc.

License: Apache License 2.0

Python 94.52% EmberScript 5.48%

rocker's People

Contributors

Stargazers

Watchers

rocker's Issues

nvidia rviz2 - is it missing --x11?

ROS 2 rviz - trying to use sommand listed, but I think it's missing --x11?
rocker --nvidia osrf/ros:crystal-desktop rviz2

combination of qt gui apps + host network crashes

I'm putting this here for reference as it was one of the first things I tried to do with rviz2. It looks like there's an issue upstream about this: moby/moby#38442

As a workaround you can use the option --noexecute and copy the docker run command adding --privileged to the invocation.

The error you will see is:

dbus[1]: The last reference on a connection was dropped without closing the connection. This is a bug in an application. See dbus_connection_unref() documentation for details.
Most likely, the application was supposed to call dbus_connection_close(), since this is a private connection.
  D-Bus not built with -rdynamic so unable to print a backtrace

Default remove container on shutdown

Is it possible to have an option to not docker run with --rm?

Distribution Not Found

Perhaps this is something to do with my system, but after installing rocker through apt on Ubuntu 18.04, I seem be getting this error:

$ rocker --nvidia --user --pull --pulse osrf/ros:crystal-desktop rviz2
Traceback (most recent call last):
  File "/usr/bin/rocker", line 6, in <module>
    from pkg_resources import load_entry_point
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 3098, in <module>
    @_call_aside
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 3082, in _call_aside
    f(*args, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 3111, in _initialize_master_working_set
    working_set = WorkingSet._build_master()
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 573, in _build_master
    ws.require(__requires__)
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 891, in require
    needed = self.resolve(parse_requirements(requirements))
  File "/usr/local/lib/python3.6/dist-packages/pkg_resources/__init__.py", line 777, in resolve
    raise DistributionNotFound(req, requirers)
pkg_resources.DistributionNotFound: The 'distro' distribution was not found and is required by rocker

This didn't seem to resolve the issue above:
https://stackoverflow.com/a/6200314

More info:

$ apt show python3-rocker 
Package: python3-rocker
Version: 0.0.1-100
Priority: optional
Section: python
Maintainer: Tully Foote <[email protected]>
Installed-Size: 48.1 kB
Depends: python3-empy, python3-pexpect, python3-requests, python3:any (>= 3.2~), python3-docker
Conflicts: python-rocker
Download-Size: 7,584 B
APT-Manual-Installed: yes
APT-Sources: http://packages.ros.org/ros/ubuntu bionic/main amd64 Packages
Description: A tool to run docker containers with customized extras
 A tool to run docker containers with customized extra added like nvidia gui support overlayed.

Feature Request: Support --name docker option

It would be great if the container started with rocker could be named so we do not have to look up random strings each time we want to exec something in the running container.

extract pull functionality into library

it's currently only available in the cli interface. It could easily be a free function for easier reuse.

Add debian package for Ubuntu 20.04 LTS (Focal Fossa)

Detect and show feedback if base images are out of date

The os_detector relies on the python base images. I've been using the locally cached version but there was a new version that i only noticed it got when I manually invoked the pull mechanism.

It would be good to at least tell users that their image is cached and out of date compared to upstream.

This relates to #86 and #26

Request for a wiki to explain various flags and potential uses

I believe this is a great tool for developing in ROS and makes the job of teaching ROS also much easier. In addition, I believe that this is a great tool for improving the quality of code that the ROS community creates.

We do need to have a wiki and at the same time also update the ROS wiki as it is very misleading and does not showcase/explain rocker's full potential.

How can I run rocker noninteractively without the container automatically stopping?

rocker/src/rocker/core.py

Line 247 in b1fae79

cmd += " -it"

This line here seems to pass -it to docker if --mode non-interactive is passed to rocker. It seems to me that actually makes the container start in interactive mode. Isn't the -d flag what we want for detached execution?

Add the ability for extensions to have requirements on other extension

For example the nvidia extension will need the x11 extension to be useful.

Also the home extension should require the user extension for permissions.

There' s a question of ordering, required before or not. And for each sequential element aka snippets.

Preamble elements are named so order independent, and arguments are order independent as far as I know as long as they're escaped properly.

This will resolve this todo:

rocker/test/test_nvidia.py

Line 134 in cea0680

 desired_plugins = ['x11', 'nvidia', 'user'] #TODO(Tfoote) encode the x11 dependency into the plugin and remove from test here 

Add a flag to pass docker parameters

It would be helpful if it was possible to pass parameters with which to call docker run. In the case of the MoveIt docker image, the rocker --user flag does not work, but a custom command would have made a fine workaround. This has come up in moveit/moveit.ros.org#282

I'm not clear on best practices here, but naively I would have imagined a call like this:

rocker --nvidia --x11 --docker-run-options '-v /home/my_user/ws_moveit:/home/root/ws_moveit' moveit/moveit:master-source bash

Might be related to #19; I'm not sure I understand the scope of that.

Enable x11 xforwarding

X based content fails when attempting to xforward as unable to connect to the x server.

root@27b685fa76ad:/# xeyes 
Error: Can't open display: localhost:10.0

It can be sent to the local display when running remotely though successfully like this

root@27b685fa76ad:/# DISPLAY=:0.0 xeyes

enable standardized validate_environment checks

Right now it's not called and done in precondition_environment usually with a sys.exit call.

Add ability to add things to the docker entrypoint using rocker extensions

Some things that are done in the overlayed image might make more sense to be supported in extending the docker entrypoint.

This would allow you to do things like extend paths, set environment variables, and other things more related to runtime. Also elements baked into the entrypoint can allow greater reuse of rocker produced images.

One possible approach would be to have a generic entrypoint that sources all setup files in a directory where extensions can insert their own elements, using a numbered file for ordering.

make robust to running against an image based off of nvidia/opengl:1.0-glvnd-devel-ubuntu16.04

Right now it crashes when it tries to reinject the nvidia settings. It would be good to either detect that and skip. Or detect that it's incompatible (such as different target platform) and error out. instead of erroring in the build step:

building >  ---> 2ffed4320c8f
building > Step 8/11 : RUN ( echo '/usr/local/lib/x86_64-linux-gnu' >> /etc/ld.so.conf.d/glvnd.conf && ldconfig || grep -q /usr/local/lib/x86_64-linux-gnu /etc/ld.so.conf.d/glvnd.conf ) &&     ( echo '/usr/local/lib/i386-linux-gnu' >> /etc/ld.so.conf.d/glvnd.conf &&  ldconfig || grep -q /usr/local/lib/i386-linux-gnu /etc/ld.so.conf.d/glvnd.conf )
building >  ---> Running in 4d3162f4f30b
building > /bin/sh: 1: cannot create /etc/ld.so.conf.d/glvnd.conf: Permission denied

building > grep: /etc/ld.so.conf.d/glvnd.conf: No such file or directory

nvidia + user on bionic fails

When extending the nvidia tests to cover xenial and bionic it failed to build on bionic using both the nvidia and user arguments.

I ended up removing the user element for the unit test as it was unnecessary: 67281fc#diff-deb510f3e7ec57637aba0ae0de5382d6L61

This appears to be a perl issue with differing versions, that's most commonly associated with mixing binary distributions and cpan distributions.

building > Step 12/14 : RUN apt-get update  && apt-get install -y     sudo  && apt-get clean
building >  ---> Using cache
building >  ---> 9d5a0b1cf12d
building > Step 13/14 : RUN useradd -U --uid 1000 -ms /bin/bash tfoote  && echo "tfoote:tfoote" | chpasswd  && adduser tfoote sudo  && echo "tfoote ALL=NOPASSWD: ALL" >> /etc/sudoers.d/tfoote
building >  ---> Running in 8458b0534fea
building > Fcntl.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)

no more output and success not detected

In a previous build I also detected an issue slightly earlier, but it seemed to recover:

Need to get 428 kB of archives.
After this operation, 1765 kB of additional disk space will be used.
Get:1 http://archive.ubuntu.com/ubuntu bionic/main amd64 sudo amd64 1.8.21p2-3ubuntu1 [428 kB]
building > IO.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)

building > Fetched 428 kB in 1s (403 kB/s)
building > Selecting previously unselected package sudo.

Getting into that container I was able to confirm that it's the adduser command that's crashing

$ docker run -ti --rm 9d5a0b1cf12d bash
root@b64960dbb3e3:/# adduser -h
Fcntl.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)

But there's nothing from cpan in the system

root@b64960dbb3e3:/# env | grep PERL
root@b64960dbb3e3:/#

Things like a dist-upgrade don't help either.

Perl itself seems to be running fine

root@b64960dbb3e3:/# perl --version

This is perl 5, version 26, subversion 1 (v5.26.1) built for x86_64-linux-gnu-thread-multi
(with 67 registered patches, see perl -V for more detail)

Copyright 1987-2017, Larry Wall

Perl may be copied only under the terms of either the Artistic License or the
GNU General Public License, which may be found in the Perl 5 source kit.

Complete documentation for Perl, including FAQ lists, should be found on
this system using "man perl" or "perldoc perl".  If you have access to the
Internet, point your browser at http://www.perl.org/, the Perl Home Page.

root@b64960dbb3e3:/# which perl
/usr/bin/perl

Other apt functions are having trouble:

root@b64960dbb3e3:/# sudo apt-get update
Hit:1 http://archive.ubuntu.com/ubuntu bionic InRelease
Get:2 http://security.ubuntu.com/ubuntu bionic-security InRelease [88.7 kB]
Get:3 http://archive.ubuntu.com/ubuntu bionic-updates InRelease [88.7 kB]
Get:4 http://archive.ubuntu.com/ubuntu bionic-backports InRelease [74.6 kB]          
Fetched 252 kB in 1s (208 kB/s)                              
Reading package lists... Done
root@b64960dbb3e3:/# sudo apt-get dist-upgrade
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Calculating upgrade... Done
The following packages will be upgraded:
  apt libapt-pkg5.0 tar
3 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
1 not fully installed or removed.
Need to get 2203 kB of archives.
After this operation, 22.5 kB of additional disk space will be used.
Do you want to continue? [Y/n] 
Get:1 http://archive.ubuntu.com/ubuntu bionic-updates/main amd64 tar amd64 1.29b-2ubuntu0.1 [234 kB]
Get:2 http://archive.ubuntu.com/ubuntu bionic-updates/main amd64 libapt-pkg5.0 amd64 1.6.8 [804 kB]
Get:3 http://archive.ubuntu.com/ubuntu bionic-updates/main amd64 apt amd64 1.6.8 [1165 kB]
Fetched 2203 kB in 2s (1399 kB/s)
IO.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
(Reading database ... 4762 files and directories currently installed.)
Preparing to unpack .../tar_1.29b-2ubuntu0.1_amd64.deb ...
Unpacking tar (1.29b-2ubuntu0.1) over (1.29b-2) ...
Setting up tar (1.29b-2ubuntu0.1) ...
(Reading database ... 4762 files and directories currently installed.)
Preparing to unpack .../libapt-pkg5.0_1.6.8_amd64.deb ...
Unpacking libapt-pkg5.0:amd64 (1.6.8) over (1.6.6ubuntu0.1) ...
Setting up libapt-pkg5.0:amd64 (1.6.8) ...
(Reading database ... 4762 files and directories currently installed.)
Preparing to unpack .../archives/apt_1.6.8_amd64.deb ...
Unpacking apt (1.6.8) over (1.6.6ubuntu0.1) ...
Setting up apt (1.6.8) ...
Installing new version of config file /etc/apt/apt.conf.d/01autoremove ...
Fcntl.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
Setting up less (487-0.1) ...
Fcntl.c: loadable library and perl binaries are mismatched (got handshake key 0xdb00080, needed 0xde00080)
dpkg: error processing package less (--configure):
 installed less package post-installation script subprocess returned error exit status 1
Processing triggers for libc-bin (2.27-3ubuntu1) ...
Errors were encountered while processing:
 less
E: Sub-process /usr/bin/dpkg returned an error code (1)

What I believe is the equivalent command line for the failing test works fine
rocker --nvidia --user testfixture_bionic_glmark2

display build errors

When there's a build error the output in rocker is suppressed.

As you can see manually running the command afterwards shows that it's failing due to executable file not found. But the rocker build just ends with an error code and no information. It is an invalid configuration but that doesn't mean we shouldn't show the error message.

$ rocker --home --user hello-world --noex
Plugins found: ['dev_helpers', 'home', 'nvidia', 'pulse', 'user']
Active extensions ['home', 'user']
Writing dockerfile to /tmp/tmpwfb8q_7_/Dockerfile
vvvvvv
# Preamble from extension [home]

# Preamble from extension [user]


FROM hello-world
# Snippet from extension [home]

# Snippet from extension [user]
# make sure sudo is installed to be able to give user sudo access in docker
RUN apt-get update \
 && apt-get install -y \
    sudo \
 && apt-get clean

RUN useradd -U --uid 1000 -ms /bin/bash osrf \
 && echo "osrf:osrf" | chpasswd \
 && adduser osrf sudo \
 && echo "osrf ALL=NOPASSWD: ALL" >> /etc/sudoers.d/osrf
# Commands below run as the developer user
USER osrf


^^^^^^
Building docker file with arguments:  {'path': '/tmp/tmpwfb8q_7_', 'rm': True, 'decode': True, 'nocache': False, 'tag': 'rocker_hello-world_home_user'}
building > Step 1/4 : FROM hello-world
building >  ---> fce289e99eb9
building > Step 2/4 : RUN apt-get update  && apt-get install -y     sudo  && apt-get clean
building >  ---> Running in 6f95c70df683
2 osrf@osrf-toughpad1:~$ docker run -ti --rm 6f95c70df683
Unable to find image '6f95c70df683:latest' locally
docker: Error response from daemon: pull access denied for 6f95c70df683, repository does not exist or may require 'docker login'.
See 'docker run --help'.
125 osrf@osrf-toughpad1:~$ docker run -ti --rm fce289e99eb9 bash
docker: Error response from daemon: OCI runtime create failed: container_linux.go:344: starting container process caused "exec: \"bash\": executable file not found in $PATH": unknown.
127 osrf@osrf-toughpad1:~$ docker run -ti --rm fce289e99eb9 sh
docker: Error response from daemon: OCI runtime create failed: container_linux.go:344: starting container process caused "exec: \"sh\": executable file not found in $PATH": unknown.
127 osrf@osrf-toughpad1:~$ docker run -ti --rm fce289e99eb9 

Hello from Docker!
This message shows that your installation appears to be working correctly.

To generate this message, Docker took the following steps:
 1. The Docker client contacted the Docker daemon.
 2. The Docker daemon pulled the "hello-world" image from the Docker Hub.
    (amd64)
 3. The Docker daemon created a new container from that image which runs the
    executable that produces the output you are currently reading.
 4. The Docker daemon streamed that output to the Docker client, which sent it
    to your terminal.

To try something more ambitious, you can run an Ubuntu container with:
 $ docker run -it ubuntu bash

Share images, automate workflows, and more with a free Docker ID:
 https://hub.docker.com/

For more examples and ideas, visit:
 https://docs.docker.com/get-started/

bump version higher than 0.1.0 to automatically override old the old rocker installations

The project on pypi has been migrated to this, but the old version was 0.1.0 so the next increment should go above that to make sure to install over the older ones.

X11 extension broken on multi-user systems

The --x11 extension fails to run xauth on the host if the hard-coded /tmp/.docker.xauth already exists and has been created by another user (permissions are 0600):

xauth:  timeout in locking authority file /tmp/.docker.xauth
Failed setting up XAuthority with command xauth nlist localhost:11.0 | sed -e 's/^..../ffff/' | xauth -f /tmp/.docker.xauth nmerge -
Failed to precondition for extension [x11] with error: Command 'xauth nlist localhost:11.0 | sed -e 's/^..../ffff/' | xauth -f /tmp/.docker.xauth nmerge -' returned non-zero exit status 1.
deactivating

I suggest to create the file with tempfile.mkstemp instead on the host, such that each rocker invocation is independent of previous runs:

rocker/src/rocker/nvidia_extension.py

Line 44 in b4a736b

self.xauth = '/tmp/.docker.xauth'

Error not raised when docker daemon not running

I get the following error when docker is not running and I try to use rocker

Traceback (most recent call last):
  File "/usr/lib/python3.9/site-packages/urllib3/connectionpool.py", line 699, in urlopen
    httplib_response = self._make_request(
  File "/usr/lib/python3.9/site-packages/urllib3/connectionpool.py", line 394, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python3.9/http/client.py", line 1255, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1301, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1250, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1010, in _send_output
    self.send(msg)
  File "/usr/lib/python3.9/http/client.py", line 950, in send
    self.connect()
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/transport/unixconn.py", line 43, in connect
    sock.connect(self.unix_socket)
ConnectionRefusedError: [Errno 111] Connection refused

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/lib/python3.9/site-packages/requests/adapters.py", line 439, in send
    resp = conn.urlopen(
  File "/usr/lib/python3.9/site-packages/urllib3/connectionpool.py", line 755, in urlopen
    retries = retries.increment(
  File "/usr/lib/python3.9/site-packages/urllib3/util/retry.py", line 531, in increment
    raise six.reraise(type(error), error, _stacktrace)
  File "/usr/lib/python3.9/site-packages/urllib3/packages/six.py", line 734, in reraise
    raise value.with_traceback(tb)
  File "/usr/lib/python3.9/site-packages/urllib3/connectionpool.py", line 699, in urlopen
    httplib_response = self._make_request(
  File "/usr/lib/python3.9/site-packages/urllib3/connectionpool.py", line 394, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python3.9/http/client.py", line 1255, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1301, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1250, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.9/http/client.py", line 1010, in _send_output
    self.send(msg)
  File "/usr/lib/python3.9/http/client.py", line 950, in send
    self.connect()
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/transport/unixconn.py", line 43, in connect
    sock.connect(self.unix_socket)
urllib3.exceptions.ProtocolError: ('Connection aborted.', ConnectionRefusedError(111, 'Connection refused'))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/api/client.py", line 214, in _retrieve_server_version
    return self.version(api_version=False)["ApiVersion"]
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/api/daemon.py", line 181, in version
    return self._result(self._get(url), json=True)
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/utils/decorators.py", line 46, in inner
    return f(self, *args, **kwargs)
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/api/client.py", line 237, in _get
    return self.get(url, **self._set_request_timeout(kwargs))
  File "/usr/lib/python3.9/site-packages/requests/sessions.py", line 555, in get
    return self.request('GET', url, **kwargs)
  File "/usr/lib/python3.9/site-packages/requests/sessions.py", line 542, in request
    resp = self.send(prep, **send_kwargs)
  File "/usr/lib/python3.9/site-packages/requests/sessions.py", line 655, in send
    r = adapter.send(request, **kwargs)
  File "/usr/lib/python3.9/site-packages/requests/adapters.py", line 498, in send
    raise ConnectionError(err, request=request)
requests.exceptions.ConnectionError: ('Connection aborted.', ConnectionRefusedError(111, 'Connection refused'))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/ridhwan/.local/bin/rocker", line 8, in <module>
    sys.exit(main())
  File "/home/ridhwan/.local/lib/python3.9/site-packages/rocker/cli.py", line 45, in main
    extension_manager.extend_cli_parser(parser, default_args)
  File "/home/ridhwan/.local/lib/python3.9/site-packages/rocker/core.py", line 99, in extend_cli_parser
    p.register_arguments(parser, default_args)
  File "/home/ridhwan/.local/lib/python3.9/site-packages/rocker/extensions.py", line 137, in register_arguments
    client = get_docker_client()
  File "/home/ridhwan/.local/lib/python3.9/site-packages/rocker/core.py", line 121, in get_docker_client
    docker_client = docker.APIClient()
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/api/client.py", line 197, in __init__
    self._version = self._retrieve_server_version()
  File "/home/ridhwan/.local/lib/python3.9/site-packages/docker/api/client.py", line 221, in _retrieve_server_version
    raise DockerException(
docker.errors.DockerException: Error while fetching server API version: ('Connection aborted.', ConnectionRefusedError(111, 'Connection refused'))

We do have a dependency missing exception but it isn't triggered here. Maybe we should add this as well.

Status of docker on my system:

Docker installed
User added to docker group
Docker daemon not started

How to reproduce:

Run rocker without starting docker daemon.

add ability to extend rocker images with rocker

rocker can generate images, for convenient sharing these images could be uploaded to a registry or the Dockerfile can be committed to a repository with auto builds enabled on dockerhub.

Then rocker could use that

A --no-build option would be valuable to support this
A way to output the Dockerfile to file would be useful.

Even more powerful would be if the extensions used in the layer were captured so that rocker run could generate the correct runtime arguments for the extensions embedded into the image. This could possibly be implemented using labels or marker files in the image. Potentially the image could embed the rocker extensions in a known location such that the extensions would not need to be installed on the host system.

This would be valuable for reproducible aggregate images. The host system verification checks could still be run.

Add ability to run things as user in the snippets

Talking with @sloretz it would be good to have the ability to do things as the user in the Dockerfile snippets.

I think that the recommended pattern would be that the snippet would switch to the USER command at the top of it and then return to the USER root at the end. To make this useful though a few helper functions would be valuable to let the plugins query for the username to switch to. Alternatively they could ask for their snippet to be run as user.

Also related the user account would need to be setup at the beginning before the snippet and the final USER setting postpended.

This would drive a need for extension point dependencies. And maybe the --user option would drive the --user-account option and anything that wants to run as a user would also require the --user-account option. And we could make sure that user-account is run first.

A slightly more integrated solution would be to allow the snippet to be registered with an option "run as user" and then the switch to and from USER would be handled by rocker instead of by the snippet. This would make it less likely that a snippet would forget to return the USER to root and break everything downstream.

Environment variable `TERM` cannot be overwritten

I use rocker --network host --device /dev/dri/card0 --x11 --pulse --user --home --env "DISPLAY" "XDG_CONFIG_HOME=~/.config" TERM=xterm-256color -- my_image. The final docker call is docker run -it --rm --device /dev/dri/card0 --device /dev/dri/card0 -e DISPLAY -e 'XDG_CONFIG_HOME=~/.config' -e TERM=xterm-256color -v /home/gael:/home/gael --network host -v /run/user/1000/pulse:/run/user/1000/pulse --device /dev/snd -e PULSE_SERVER=unix:/run/user/1000/pulse/native -v /run/user/1000/pulse/native:/run/user/1000/pulse/native --group-add 29 -e DISPLAY -e TERM -e QT_X11_NO_MITSHM=1 -e XAUTHORITY=/tmp/.docker.xauth -v /tmp/.docker.xauth:/tmp/.docker.xauth -v /tmp/.X11-unix:/tmp/.X11-unix -v /etc/localtime:/etc/localtime:ro d3ddfa229ca8.

The TERM environment is set twice and the last one takes TERM from the calling environment, which in my case doesn't work because this particular terminal is not installed in the image. I know some workarounds but this is inconvenient.

pass through window resize logic to interactive window

By default most things go through except for the window resize signals.

https://pexpect.readthedocs.io/en/stable/api/pexpect.html#pexpect.spawn.interact

Nvidia compatibilty not working with focal host and xenial image

On my Focal host I cannot run the nvidia images on xenial 16.04 anymore. However the tests pass on my bionic host 18.04.

The nvidia 450 driver and the nvidia 440 drivers are respectively installed.

focal install missing python3-distutils by default

The quick solution is to sudo apt install python3-distutils but it should be declared properly.

Update user option to use runtime user fixing

https://github.com/boxboat/fixuid uses a Go program to change the UID and GID inside, we could leverage this or use the same technique with a script bash or python.

platform detection breaks nvidia plugin if the entrypoint is invalid

This would be more robust if it provided it's own entrypoint.

Backtrace of the error.

$ rocker --x11 --nvidia --user --home --pulse ddm
Plugins found: ['dev_helpers', 'home', 'nvidia', 'pulse', 'user', 'x11']
Active extensions ['home', 'nvidia', 'pulse', 'x11', 'user']
Traceback (most recent call last):
  File "/usr/bin/rocker", line 11, in <module>
    load_entry_point('rocker==0.1.4', 'console_scripts', 'rocker')()
  File "/usr/lib/python3/dist-packages/rocker/cli.py", line 58, in main
    dig = DockerImageGenerator(active_extensions, args_dict, base_image)
  File "/usr/lib/python3/dist-packages/rocker/core.py", line 77, in __init__
    self.dockerfile = generate_dockerfile(active_extensions, self.cliargs, base_image)
  File "/usr/lib/python3/dist-packages/rocker/core.py", line 180, in generate_dockerfile
    dockerfile_str += el.get_preamble(args_dict) + '\n'
  File "/usr/lib/python3/dist-packages/rocker/nvidia_extension.py", line 107, in get_preamble
    return em.expand(preamble, self.get_environment_subs(cliargs))
  File "/usr/lib/python3/dist-packages/rocker/nvidia_extension.py", line 92, in get_environment_subs
    dist, ver, codename = detect_os(cliargs['base_image'])
TypeError: 'NoneType' object is not iterable

Example of the error that caused the issue:

docker run -it   --rm    -v /home/tfoote:/home/tfoote  -v /run/user/1000/pulse:/run/user/1000/pulse --device /dev/snd  -e PULSE_SERVER=unix:/run/user/1000/pulse/native -v /run/user/1000/pulse/native:/run/user/1000/pulse/native --group-add 29   -e DISPLAY -e TERM   -e QT_X11_NO_MITSHM=1   -e XAUTHORITY=/tmp/.docker.xauth -v /tmp/.docker.xauth:/tmp/.docker.xauth   -v /tmp/.X11-unix:/tmp/.X11-unix   -v /etc/localtime:/etc/localtime:ro    rocker_ddm_home_pulse_x11_user --entrypoint /bin/bash
/entrypoint.sh: 4: .: Can't open /workspace/drone_demo/install/setup.sh

The 'docker' distribution was not found and is required by rocker

Traceback (most recent call last):
File "/usr/bin/rocker", line 6, in
from pkg_resources import load_entry_point
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 2927, in
@_call_aside
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 2913, in _call_aside
f(*args, **kwargs)
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 2940, in _initialize_master_working_set
working_set = WorkingSet._build_master()
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 635, in _build_master
ws.require(requires)
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 943, in require
needed = self.resolve(parse_requirements(requirements))
File "/usr/lib/python3/dist-packages/pkg_resources/init.py", line 829, in resolve
raise DistributionNotFound(req, requirers)
pkg_resources.DistributionNotFound: The 'docker' distribution was not found and is required by rocker

detect missing docker permissions and provide a useful prompt

Right now you get a long traceback if you don't have permissions to build the docker container

$ rocker docker/hello-world                                                    
Plugins found: ['dev_helpers', 'home', 'nvidia', 'pulse', 'user']
Active extensions []                                                           
Writing dockerfile to /tmp/tmp93bgqbui/Dockerfile
vvvvvv                                                                                                

FROM docker/hello-world

^^^^^^
Building docker file with arguments:  {'path': '/tmp/tmp93bgqbui', 'rm': True, 'decode': True, 'nocache': False, 'tag': 'rocker_docker/hello-world'}
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 601, in urlopen
    chunked=chunked)
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 357, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python3.6/http/client.py", line 1239, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1285, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1234, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1026, in _send_output
    self.send(msg)
  File "/usr/lib/python3.6/http/client.py", line 964, in send
    self.connect()
  File "/usr/lib/python3/dist-packages/docker/transport/unixconn.py", line 33, in connect
    sock.connect(self.unix_socket)
PermissionError: [Errno 13] Permission denied

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/requests/adapters.py", line 440, in send
    timeout=timeout
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 639, in urlopen
    _stacktrace=sys.exc_info()[2])
  File "/usr/lib/python3/dist-packages/urllib3/util/retry.py", line 357, in increment
    raise six.reraise(type(error), error, _stacktrace)
  File "/usr/lib/python3/dist-packages/six.py", line 692, in reraise
    raise value.with_traceback(tb)
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 601, in urlopen
    chunked=chunked)
  File "/usr/lib/python3/dist-packages/urllib3/connectionpool.py", line 357, in _make_request
    conn.request(method, url, **httplib_request_kw)
  File "/usr/lib/python3.6/http/client.py", line 1239, in request
    self._send_request(method, url, body, headers, encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1285, in _send_request
    self.endheaders(body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1234, in endheaders
    self._send_output(message_body, encode_chunked=encode_chunked)
  File "/usr/lib/python3.6/http/client.py", line 1026, in _send_output
    self.send(msg)
  File "/usr/lib/python3.6/http/client.py", line 964, in send
    self.connect()
  File "/usr/lib/python3/dist-packages/docker/transport/unixconn.py", line 33, in connect
    sock.connect(self.unix_socket)
urllib3.exceptions.ProtocolError: ('Connection aborted.', PermissionError(13, 'Permission denied'))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/bin/rocker", line 11, in <module>
    load_entry_point('rocker==0.1.3', 'console_scripts', 'rocker')()
  File "/usr/lib/python3/dist-packages/rocker/cli.py", line 63, in main
    exit_code = dig.build(**vars(args))
  File "/usr/lib/python3/dist-packages/rocker/core.py", line 64, in build
    for line in docker_client.build(**arguments):
  File "/usr/lib/python3/dist-packages/docker/api/build.py", line 246, in build
    timeout=timeout,
  File "/usr/lib/python3/dist-packages/docker/utils/decorators.py", line 46, in inner
    return f(self, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/docker/api/client.py", line 185, in _post
    return self.post(url, **self._set_request_timeout(kwargs))
  File "/usr/lib/python3/dist-packages/requests/sessions.py", line 567, in post
    return self.request('POST', url, data=data, json=json, **kwargs)
  File "/usr/lib/python3/dist-packages/requests/sessions.py", line 520, in request
    resp = self.send(prep, **send_kwargs)
  File "/usr/lib/python3/dist-packages/requests/sessions.py", line 630, in send
    r = adapter.send(request, **kwargs)
  File "/usr/lib/python3/dist-packages/requests/adapters.py", line 490, in send
    raise ConnectionError(err, request=request)
requests.exceptions.ConnectionError: ('Connection aborted.', PermissionError(13, 'Permission denied'))

No title bar on Gazebo under Ubuntu 20

Hi,

I tried to start Gazebo 7.0 under Ubuntu 20 by running the command in the README.file. However, there is no title bar on the top and I cannot resize this window.

Could anyone help me?

Break build cache only when baseimage has changed

For every invocation of something like rocker --x11 --nvidia .. , it seems that rocker rebuilds the same dockerfile with no caching, even if the base image/tag referred to in command remains unchanged. The results in slow spin up of containers as well as network access requirements, as users must wait apt install steps to complete. I think rocker should only seek to rebuild the child image used for spawning the eventual container when the child image/tag is out of sync with the parent image/tag provided. So, what is the current source of entropy that is breaking the docker build cache?

Check for minimum version of docker-py necessary

Users have run into issues which is likely too low a version on xenial: osrf/car_demo#49 (comment)

[nvidia extension] Detect an unsupported nvidia driver

You can get errors like:

/usr/bin/docker: Error response from daemon: could not select device driver "" with capabilities: [[gpu]]

If you don't have a new enough nvidia driver.

Use docker instead of docker-py

The docker-py library was renamed to docker in 2.0.0 https://github.com/docker/docker-py/releases/tag/2.0.0 . If both docker-py and docker are installed then import docker is the module that was found first.

Show some sort of feedback when downloading docker images

It can be confusing when rocker appears to be doing nothing, but it's actually downloading the docker images necessary to run.

Getting hooks into the actual download might be a challenge, but if we detect and prefetch the base images before hand we can provide feedback during that and not have to worry about getting the feedback from the build stage.

This likely will overlap with #26

Also make sure to cover intermediate/helper image fetching such as the python image used for the os detector. It's 900Mb which is bigger than I'd hoped.

Can I use my own image ?

rocker does not work with rootless docker

I have installed docker rootless on a machine of our computer pool for students. We do not want the students to have root-like access, thus we avoid setting up the docker group.

When trying a simple rocker --nvidia nvidia/cuda:11.0-base nvidia-smi, the following error occurs:

rocker: error: DependencyMissing encountered: Docker Client failed to connect to docker daemon. Please verify that docker is installed and running. As well as that you have permission to access the docker daemon. This is usually by being a member of the docker group.

The error occurs when instantiating the docker.APIClient() which can not find the rootless docker instance. However, docker.from_env().api works and could serve as a drop in replacement without any other code changes. I will open a pull request, when I'm done testing my container.

On the nvidia side, I also had to change the /etc/nvidia-container-runtime/config.toml to allow the device access.

[nvidia-container-cli]
no-cgroups = true

create empty home directory from skeleton if --user enabled without --home

Otherwise programs that try to use the home directory will get permission denied errors.

Such as osrf/car_demo#56

Resources:

https://askubuntu.com/questions/152707/how-to-make-user-home-folder-after-account-creation
https://askubuntu.com/questions/335961/create-default-home-directory-for-existing-user-in-terminal

It looks like mkhomedir_helper might be available widely enough.
http://man7.org/linux/man-pages/man8/mkhomedir_helper.8.html

nvidia overlay failing on bionic: x86_64-linux-gnu: no such file or directory

The auto generated Dockerfile seems to encounter a build issue in the first stage of it multistage build.

$ rocker --nvidia --user --pull --pulse osrf/ros:crystal-desktop rviz2
Plugins found: ['dev_helpers', 'home', 'nvidia', 'pulse', 'user']
Active extensions ['nvidia', 'pulse', 'user']
Pulling image osrf/ros:crystal-desktop
b'{"status":"Pulling from osrf/ros","id":"crystal-desktop"}\r\n'
b'{"status":"Digest: sha256:62579d806386b2b3ff2ecf3924f0ba39b8d8422fb7979db42e469cfbfd26e19f"}\r\n'
b'{"status":"Status: Image is up to date for osrf/ros:crystal-desktop"}\r\n'
Writing dockerfile to /tmp/tmpg_lv8g3_/Dockerfile
vvvvvv
...
^^^^^^
Building docker file with arguments:  {'path': '/tmp/tmpg_lv8g3_', 'rm': True, 'decode': True, 'nocache': False, 'tag': 'rocker_osrf/ros:crystal-desktop_nvidia_pulse_user'}
building > Step 1/16 : FROM nvidia/opengl:1.0-glvnd-devel-ubuntu18.04 as glvnd
building >  ---> 04c47f164887
building > Step 2/16 : FROM osrf/ros:crystal-desktop
building >  ---> c3d97d8c2f50
building > Step 3/16 : COPY --from=glvnd /usr/local/lib/x86_64-linux-gnu /usr/local/lib/x86_64-linux-gnu
Cannot run if build has not passed.

Would be nice if the rocker could forward the build error to the printout, rather than needing to manually build the same dockerfile to identify the exact error.

$ docker build -t temp /tmp/tmpg_lv8g3_/
Sending build context to Docker daemon  4.096kB
Step 1/16 : FROM nvidia/opengl:1.0-glvnd-devel-ubuntu18.04 as glvnd
 ---> 04c47f164887
Step 2/16 : FROM osrf/ros:crystal-desktop
 ---> c3d97d8c2f50
Step 3/16 : COPY --from=glvnd /usr/local/lib/x86_64-linux-gnu /usr/local/lib/x86_64-linux-gnu
COPY failed: stat /var/lib/docker/overlay2/471568f79754e1b366acbc4647bda7de3ccc8bde738d60023154c746cbcb85a7/merged/usr/local/lib/x86_64-linux-gnu: no such file or directory

Dockerfile

# Preamble from extension [nvidia]
# Ubuntu 16.04 with nvidia-docker2 beta opengl support
FROM nvidia/opengl:1.0-glvnd-devel-ubuntu18.04 as glvnd

# Preamble from extension [pulse]

# Preamble from extension [user]


FROM osrf/ros:crystal-desktop
# Snippet from extension [nvidia]
# Open nvidia-docker2 GL support
COPY --from=glvnd /usr/local/lib/x86_64-linux-gnu /usr/local/lib/x86_64-linux-gnu
COPY --from=glvnd /usr/local/lib/i386-linux-gnu /usr/local/lib/i386-linux-gnu
COPY --from=glvnd /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu
COPY --from=glvnd /usr/lib/i386-linux-gnu /usr/lib/i386-linux-gnu

COPY --from=glvnd /usr/local/share/glvnd/egl_vendor.d/10_nvidia.json /usr/local/share/glvnd/egl_vendor.d/10_nvidia.json

# if the path is alreaady present don't fail because of being unable to append
RUN ( echo '/usr/local/lib/x86_64-linux-gnu' >> /etc/ld.so.conf.d/glvnd.conf && ldconfig || grep -q /usr/local/lib/x86_64-linux-gnu /etc/ld.so.conf.d/glvnd.conf ) && \
    ( echo '/usr/local/lib/i386-linux-gnu' >> /etc/ld.so.conf.d/glvnd.conf &&  ldconfig || grep -q /usr/local/lib/i386-linux-gnu /etc/ld.so.conf.d/glvnd.conf )

ENV LD_LIBRARY_PATH /usr/local/lib/x86_64-linux-gnu:/usr/local/lib/i386-linux-gnu${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

ENV NVIDIA_VISIBLE_DEVICES ${NVIDIA_VISIBLE_DEVICES:-all}
ENV NVIDIA_DRIVER_CAPABILITIES ${NVIDIA_DRIVER_CAPABILITIES:+$NVIDIA_DRIVER_CAPABILITIES,}graphics

# Snippet from extension [pulse]
RUN mkdir -p /etc/pulse
RUN echo '\n\
# Connect to the hosts server using the mounted UNIX socket\n\
default-server = unix:/run/user/1000/pulse/native\n\
\n\
# Prevent a server running in the container\n\
autospawn = no\n\
daemon-binary = /bin/true\n\
\n\
# Prevent the use of shared memory\n\
enable-shm = false\n\
\n'\
> /etc/pulse/client.conf

# Snippet from extension [user]
# make sure sudo is installed to be able to give user sudo access in docker
RUN apt-get update \
 && apt-get install -y \
    sudo \
 && apt-get clean


RUN useradd -U --uid 1000 -ms /bin/bash ubuntu \
 && echo "ubuntu:ubuntu" | chpasswd \
 && adduser ubuntu sudo \
 && echo "ubuntu ALL=NOPASSWD: ALL" >> /etc/sudoers.d/ubuntu

# Commands below run as the developer user
USER ubuntu

tests crash when run without displaying the console output

It seems that the console output proxy in empy breaks.

======================================================================
ERROR: test_nvidia_glmark2 (test_nvidia.NvidiaTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2686, in installProxy
    sys.stdout._testProxy()
AttributeError: '_io.StringIO' object has no attribute '_testProxy'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/tfoote/work/github/tfoote/rocker/test/test_nvidia.py", line 52, in test_nvidia_glmark2
    dig = DockerImageGenerator(active_extensions, '', self.dockerfile_tag)
  File "/home/tfoote/work/github/tfoote/rocker/src/rocker/core.py", line 38, in __init__
    self.dockerfile = generate_dockerfile(active_extensions, self.cliargs, base_image)
  File "/home/tfoote/work/github/tfoote/rocker/src/rocker/core.py", line 140, in generate_dockerfile
    dockerfile_str += el.get_preamble(args_dict) + '\n'
  File "/home/tfoote/work/github/tfoote/rocker/src/rocker/extensions.py", line 120, in get_preamble
    return em.expand(preamble, self.get_environment_subs())
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 3028, in expand
    globals=_globals)
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2078, in __init__
    self.installProxy()
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2693, in installProxy
    raise Error("interpreter stdout proxy lost")
em.Error: interpreter stdout proxy lost

Exception ignored in: <bound method Interpreter.__del__ of <empy pseudomodule/interpreter at 0x7fbf694cac50>>
Traceback (most recent call last):
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2094, in __del__
    self.shutdown()
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2157, in shutdown
    self.finalize()
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2646, in finalize
    self.push()
  File "/tmp/test_venv/lib/python3.5/site-packages/empy-3.3.2-py3.5.egg/em.py", line 2200, in push
    sys.stdout.push(self)
AttributeError: '_io.StringIO' object has no attribute 'push'

The short term workaround is to just invoke nosetests with the -s option

Ubuntu 20.04 support for `--nvidia`

Currently testing some stuff with ROS2 Foxy on Ubutu 20.04. Looks like --nvidia isn't supported yet:

WARNING distro version 20.04 not in supported list by Nvidia supported versions ['16.04', '18.04']

Tested on OSX/MacOS

Does anyone have experience testing/using this on MacOS?

Name of the package may be a bit confusing

Hello, thanks for sharing this package!

I'm just concerned about people googling for it will most probably end up in:

https://github.com/grammarly/rocker

As it was a relatively famous (and very useful) package, now discontinued.

issue with running drone 'mount path must be absolute'

I thought I had this running at least once today with no changes and then ran into some issues where now it is saying i need an absolute path:

docker command:

sudo rocker --nvidia --x11 --user --home --pull --pulse tfoote/drone_demo

Executing command: 
docker run -it   --rm    -v /home/roboubuntu:/home/roboubuntu   --gpus all -v /run/user/0/pulse:/run/user/0/pulse --device /dev/snd  -e PULSE_SERVER=unix:None/pulse/native -v None/pulse/native:None/pulse/native --group-add 29   -e DISPLAY -e TERM   -e QT_X11_NO_MITSHM=1   -e XAUTHORITY=/tmp/.docker.xauth -v /tmp/.docker.xauth:/tmp/.docker.xauth   -v /tmp/.X11-unix:/tmp/.X11-unix   -v /etc/localtime:/etc/localtime:ro    38a84f5027c9 
/usr/bin/docker: Error response from daemon: invalid volume specification: 'None/pulse/native:None/pulse/native': invalid mount config for type "volume": invalid mount path: 'None/pulse/native' mount path must be absolute.
See '/usr/bin/docker run --help'.

intermittent pexpect failures

I ran rocker inside byobu and got the following traceback.

Immediately rerunning it did not cause the same error.

It appeared to be hung but this may be it downloading a bigger image in the background because it was a fresh machine. And now with the silent pull in the background pexpect might timeout due to no output from the process. If confirmed this will be related to #86

Successfully built e02c380413c7
running,  docker run -it --rm e02c380413c7
Traceback (most recent call last):
  File "/home/ubuntu/venv/bin/rocker", line 11, in <module>
    load_entry_point('rocker==0.2.3', 'console_scripts', 'rocker')()
  File "/home/ubuntu/venv/lib/python3.8/site-packages/rocker-0.2.3-py3.8.egg/rocker/cli.py", line 72, in main
  File "/home/ubuntu/venv/lib/python3.8/site-packages/rocker-0.2.3-py3.8.egg/rocker/core.py", line 249, in run
  File "/home/ubuntu/novnc-rocker/novnc_rocker/turbovnc.py", line 19, in precondition_environment
    detected_os = detect_os(cli_args['base_image'], print, nocache=cli_args.get('nocache', False))
  File "/home/ubuntu/venv/lib/python3.8/site-packages/rocker-0.2.3-py3.8.egg/rocker/os_detector.py", line 65, in detect_os
  File "/home/ubuntu/venv/lib/python3.8/site-packages/pexpect-4.8.0-py3.8.egg/pexpect/spawnbase.py", line 444, in read
    self.expect(self.delimiter)
  File "/home/ubuntu/venv/lib/python3.8/site-packages/pexpect-4.8.0-py3.8.egg/pexpect/spawnbase.py", line 343, in expect
    return self.expect_list(compiled_pattern_list,
  File "/home/ubuntu/venv/lib/python3.8/site-packages/pexpect-4.8.0-py3.8.egg/pexpect/spawnbase.py", line 372, in expect_list
    return exp.expect_loop(timeout)
  File "/home/ubuntu/venv/lib/python3.8/site-packages/pexpect-4.8.0-py3.8.egg/pexpect/expect.py", line 181, in expect_loop
    return self.timeout(e)
  File "/home/ubuntu/venv/lib/python3.8/site-packages/pexpect-4.8.0-py3.8.egg/pexpect/expect.py", line 144, in timeout
    raise exc
pexpect.exceptions.TIMEOUT: Timeout exceeded.
<pexpect.pty_spawn.spawn object at 0x7fd297ef8b20>
command: /usr/bin/docker
args: ['/usr/bin/docker', 'run', '-it', '--rm', 'e02c380413c7']
buffer (last 100 chars): b''
before (last 100 chars): ''
after: <class 'pexpect.exceptions.TIMEOUT'>
match: None
match_index: None
exitstatus: None
flag_eof: False
pid: 46996
child_fd: 5
closed: False
timeout: 30
delimiter: <class 'pexpect.exceptions.EOF'>
logfile: None
logfile_read: None
logfile_send: None
maxread: 2000
ignorecase: False
searchwindowsize: None
delaybeforesend: 0.05
delayafterclose: 0.1
delayafterterminate: 0.1
searcher: searcher_re:
    0: EOF

os detection failing due to paths

Example output output: [6] Cannot open self /tmp/staticx-OdKefo/detect_os or archive /tmp/staticx-OdKefo/detect_os.pkg

There seems to be a recent regression in PyInstaller that's compounded with staticx

Regression: pyinstaller/pyinstaller#5257 with Fix pyinstaller/pyinstaller#5258 but there's some related traffic on an older ticket: pyinstaller/pyinstaller#2357

I have not found a version of pyinstaller to pin it to that will work. It might be related to needing to pin the bootloader too. pyinstaller/pyinstaller#2357 (comment)

This does specifically work with the pyinstaller output but not after being run through staticx

There's a related older ticket here too: JonathonReinhart/staticx#71

This fails in my binary installation from debs, but I cannot reproduce it in my workspace

$ rocker --nvidia --x11 --user --home --pulse --nocache tfoote/drone_demo 
tfoote@snowman5:~ Last: [130] (0s Seconds)
$ rocker --version
rocker 0.2.2
tfoote@snowman5:~ Last: [0] (0s Seconds)
$ which rocker
/usr/bin/rocker

There is an issue that --nocache doesn't pass through to the detector.

Network options are hardcoded

I feel that the --network arg should be using the exact string passed to it to resolve which docker network to attach to, and not assume how the networks are named on the current system.

rocker/src/rocker/cli.py

Line 37 in 8f0da1f

parser.add_argument('--network', choices=['bridge', 'host', 'overlay', 'none'])

detect a base image doesn't exist and give a good error message

A good user message for this would be good. It will need to be interleaved with the pull or implicit pull appropriately.

This will also avoid the plugins potentially trying to introspect a non-existent image. Currently if nvidia is enabled it will be the first thing to crash. But if we make this catch earlier the extensions won't need to be robust.

osrf / rocker Goto Github PK

rocker's People

Contributors

Stargazers

Watchers

Forkers

rocker's Issues

Status of docker on my system:

How to reproduce:

Dockerfile

Recommend Projects

Recommend Topics

Recommend Org

Jobs