[oe] Kernel modules being built, but not being included in image

Mark Asselstine mark.asselstine at windriver.com
Wed Jan 16 02:54:32 UTC 2019 

On Tuesday, January 15, 2019 5:57:34 PM EST Mark Asselstine wrote:
> On Tuesday, January 15, 2019 1:26:31 PM EST Muhlenkamp, Lewis wrote:
> > Mark,
> > 
> > I got the docker daemon to start up.  I figured out what I needed to put
> > into my .../recipes-kernel/linux/linux-intel/docker.cfg file to make sure
> > all of the kernel modules were builts.  I also found that just adding
> > 
> >     CORE_IMAGE_EXTRA_INSTALL += "kernel-modules docker"
> > 
> > will add all of the kernel modules as well as the docker software.
> > 
> > So, the docker daemon is running.  I can run docker commands like "docker
> > image ls"  I can build a docker image, but I cannot run the docker image.
> > For example, I created a simple dockerfile that just has the line "FROM
> > ubuntu".  I can successfully run "docker build ." in that directory.  It
> > creates an image.  I can run "docker image ls" and see the image.  When I
> > run
> > 
> >     docker run -I -t ubuntu "/bin/bash"
> > 
> > though, I get an error message
> > 
> > docker: Error response from daemon: failed to create endpoint
> > elated_aryabhatadoc on network bridge; failed to add the host
> > (veth3befa72)
> > <+> sandbox (veth40a3e1c) pair interfaces: operation not supported.
> 
> Not surprisingly a container failed to start because of networking :). I
> would guess that 90% of the time I have a container fail to start it is
> networking related.
> 
> The veth pair should be independent from your network interface, either real
> on real hw or virtual in your case. Whenever you see a veth or a pair of
> veth interfaces just visualize the classic blue cat5 cable with two ends to
> it, where each end is plugged into "something". This is essentially what
> the veth pair represents. By default, when you run docker as you have the
> one end becomes the interface inside of the container, the other end is
> 'plugged' into the docker bridge on the host. Based on the message I am
> guessing the veth which is supposed to be 'plugged' into the docker bridge
> has failed to do so.
> 
> My day was a bit messed up so I only got a build put together now so
> unfortunately I won't be able to get anything more helpful put together
> until tomorrow sometime. In the mean time ensure the docker0 bridge is up
> and available. And if not figure out why it is not.


I was able to validate things. First here are my changes after I source oe-
init-build-env.

bblayers.conf
---
  /home/masselst/git/poky/layers/meta-virtualization \
  /home/masselst/git/poky/layers/meta-openembedded/meta-oe \
  /home/masselst/git/poky/layers/meta-openembedded/meta-networking \
  /home/masselst/git/poky/layers/meta-openembedded/meta-filesystems \
  /home/masselst/git/poky/layers/meta-openembedded/meta-python \


local.conf
---
DISTRO_FEATURES_append = " virtualization"
IMAGE_INSTALL_append = "docker \
 kernel-module-xt-conntrack \
 kernel-module-nf-nat \
 kernel-module-xt-addrtype"

KERNEL_MODULE_AUTOLOAD += "xt_conntrack"
KERNEL_MODULE_AUTOLOAD += "xt_addrtype"

DISTRO_FEATURES_append = " systemd"
DISTRO_FEATURES_BACKFILL_CONSIDERED += "sysvinit"
VIRTUAL-RUNTIME_init_manager = "systemd"
VIRTUAL-RUNTIME_initscripts = "systemd-compat-units"

I also set the machine to "qemux86-64".


After using 'docker pull hello-world' I run it
---
root at qemux86-64:~# docker run -it hello-world
[  434.530556] docker0: port 1(veth2c4037e) entered blocking state
[  434.533810] docker0: port 1(veth2c4037e) entered disabled state
[  434.538888] device veth2c4037e entered promiscuous mode
[  434.552937] IPv6: ADDRCONF(NETDEV_UP): veth2c4037e: link is not ready
[  435.645525] eth0: renamed from veth6909d78
[  435.648422] IPv6: ADDRCONF(NETDEV_CHANGE): veth2c4037e: link becomes ready
[  435.649944] docker0: port 1(veth2c4037e) entered blocking state
[  435.651224] docker0: port 1(veth2c4037e) entered forwarding state
[  435.654061] IPv6: ADDRCONF(NETDEV_CHANGE): docker0: link becomes ready

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/

[  435.899076] docker0: port 1(veth2c4037e) entered disabled state
[  435.904659] veth6909d78: renamed from eth0
[  435.960366] docker0: port 1(veth2c4037e) entered disabled state
[  435.967155] device veth2c4037e left promiscuous mode
[  435.970292] docker0: port 1(veth2c4037e) entered disabled state
---
I am running it from the console so you get the pollution of the veth being 
created and torn down but in this case you might be interested in seeing this 
so it is good that it is captured here.

Anyways, take a look at the above, see how it maps for you and if you continue 
to have issues I can try to assist more. You might want to run things such 
that you get more logs. I am pretty busy but possibly put your image somewhere 
and if I get the chance I can have a look.

MarkA


> 
> MarkA
> 
> > I have been doing all of my testing on a VirtualBox VM.  I'm not sure if
> > there is something missing in VirtualBox that may be causing this, or some
> > VM setting that's not properly configured.  I'm going to try on physical
> > hardware as well to see if that fixes the issue.
> > 
> > If there is something that I am missing though within my openembedded
> > build
> > that will fix this, please let me know.
> > 
> > Thank you
> > 
> > Lewis Muhlenkamp
> > 
> > -----Original Message-----
> > From: Mark Asselstine <mark.asselstine at windriver.com>
> > Sent: Monday, January 14, 2019 4:56 PM
> > To: Muhlenkamp, Lewis <lewis.muhlenkamp at stryker.com>
> > Cc: openembedded-devel <openembedded-devel at lists.openembedded.org>
> > Subject: Re: [oe] Kernel modules being built, but not being included in
> > image
> > 
> > This has been asked in the past and I did have a "mini" layer that could
> > be
> > used in addition to meta-virt to allow you to get what you need in an
> > image
> > fairly easily. We were going to do some work to make this easier but I
> > haven't looked in a while so I can't say where things are at off the top
> > of
> > my head. At any rate I am just back from some travel but I will try to
> > take
> > a look at this tomorrow, after which I should be able to provide some
> > better guidance.
> > 
> > Mark
> > On Fri, Jan 11, 2019 at 7:39 PM Muhlenkamp, Lewis
> > 
> > <lewis.muhlenkamp at stryker.com> wrote:
> > > Hello,
> > > 
> > > 
> > > 
> > > TLDR: How do I get docker fully functional in my openembedded linux
> > > image?
> > > 
> > > 
> > > 
> > > I've been trying to get docker included into my image.  All of my
> > > attempts
> > > lead to the same error messages appearing in the log file, and docker
> > > not
> > > starting.
> > > 
> > > 
> > > 
> > > The error messages are
> > > 
> > > 
> > > 
> > > === Start docker messages ===
> > > Jan 10 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.414778299Z" level=error msg="Failed to
> > > built-in
> > > GetDriver graph btrfs /var/lib/docker"
> 
>  Jan 10 15:56:25 intel-corei7-64
> 
> > > dockerd[210]: time="2019-01-10T15:56:25.460695720Z" level=warning
> > > msg="Your kernel does not support cgroup cfs period" Jan 10 15:56:25
> > > intel-corei7-64 dockerd[210]: time="2019-01-10T15:56:25.460795185Z"
> > > level=warning msg="Your kernel does not support cgroup cfs quotas" Jan
> > > 10
> > > 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.460896539Z" level=warning msg="Your kernel
> > > does
> > > not support cgroup cfs blkio weight" Jan 10 15:56:25 intel-corei7-64
> > > dockerd[210]: time="2019-01-10T15:56:25.461255643Z" level=warning
> > > msg="Your kernel does not support cgroup cfs blkio
> > > throttle.read_bps_device" Jan 10 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.461381616Z" level=warning msg="Your kernel
> > > does
> > > not support cgroup cfs blkio throttle.write_bps_device" Jan 10 15:56:25
> > > intel-corei7-64 dockerd[210]: time="2019-01-10T15:56:25.461503746Z"
> > > level=warning msg="Your kernel does not support cgroup cfs blkio
> > > throttle.read_iops_device" Jan 10 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.461601879Z" level=warning msg="Your kernel
> > > does
> > > not support cgroup cfs blkio throttle.write_iops_device" Jan 10 15:56:25
> > > intel-corei7-64 dockerd[210]: time="2019-01-10T15:56:25.475747665Z"
> > > level=warning msg="Running modprobe bridge br_netfilter failed with
> > > message: modprobe: WARNING: Module br_netfilter not found in directory
> > > /lib/modules/4.14.78-intel-pk-standard\ninsmod
> > > /lib/modules/4.14.78-intel-pk-standard/kernel/net/llc.ko \ninsmod
> > > /lib/modules/4.14.78-intel-pk-standard/kernel/net/802/stp.ko \ninsmod
> > > /lib/modules/4.14.78-intel-pk-standard/kernel/net/bridge/bridge.ko \n,
> > > error: exit status 1" Jan 10 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.659844723Z" level=warning msg="Could not load
> > > necessary modules for IPSEC rules: Running modprobe xfrm_user failed
> > > with
> > > message: `modprobe: WARNING: Module xfrm_user not found in directory
> > > /lib/modules/4.14.78-intel-pk-standard`, error: exit status 1" Jan 10
> > > 15:56:25 intel-corei7-64 dockerd[210]:
> > > time="2019-01-10T15:56:25.662494167Z" level=warning msg="Could not load
> > > necessary modules for Conntrack: Running modprobe nf_conntrack_netlink
> > > failed with message: `modprobe: WARNING: Module nf_conntrack_netlink not
> > > found in directory /lib/modules/4.14.78-intel-pk-standard`, error: exit
> > > status 1" Jan 10 15:56:25 intel-corei7-64 dockerd[210]: failed to start
> > > daemon: Error initializing network controller: Error creating default
> > > "bridge" network: Failed to program NAT chain: Failed to inject DOCKER
> > > in
> > > PREROUTING chain: iptables failed: iptables -wait -t nat -A PREROUTING
> > > -m
> > > addrtype -dst-type LOCAL -j DOCKER: iptables: No chain/target/match by
> > > that name. === End docker messages ===
> > > 
> > > 
> > > 
> > > I was using my own custom image type, but I got the same results when
> > > trying to build and use core-image-minimal.
> > > 
> > > 
> > > 
> > > I tried including the
> > > meta-virtualization/recipes/kernel/linux/linux-yocto/docker.scc stuff
> > > in,
> > > but since I set MACHINE to intel-corei7-64, I copied the docker.scc and
> > > docker.cfg into my custom layer:
> > > 
> > > 
> > > 
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/docker.cfg
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/kernel_baseline.s
> > > cc meta-stryker/common/recipes-kernel/linux/linux-intel_%.bbappend
> > > 
> > > 
> > > 
> > > That didn't seem to work either.  The modules always got built.  For
> > > example, br_netfilter.ko is built:
> > > 
> > > 
> > > 
> > > lmuhlenkamp at c71703b3ba7d:~/build-20181213a/tmp-glibc$ find . -name
> > > br_netfilter.ko
> > > ./work/corei7-64-intel-common-oe-linux/linux-intel/4.14.78+gitAUTOINC+
> > > 6a3254e7b3_56f15146cf-r0/image/lib/modules/4.14.78-intel-pk-standard/k
> > > ernel/net/bridge/br_netfilter.ko
> > > ./work/corei7-64-intel-common-oe-linux/linux-intel/4.14.78+gitAUTOINC+
> > > 6a3254e7b3_56f15146cf-r0/packages-split/kernel-module-br-netfilter-4.1
> > > 4.78-intel-pk-standard/lib/modules/4.14.78-intel-pk-standard/kernel/ne
> > > t/bridge/br_netfilter.ko
> > > ./work/corei7-64-intel-common-oe-linux/linux-intel/4.14.78+gitAUTOINC+
> > > 6a3254e7b3_56f15146cf-r0/linux-corei7-64-intel-common-standard-build/n
> > > et/bridge/br_netfilter.ko
> > > ./work/corei7-64-intel-common-oe-linux/linux-intel/4.14.78+gitAUTOINC+
> > > 6a3254e7b3_56f15146cf-r0/package/lib/modules/4.14.78-intel-pk-standard
> > > /kernel/net/bridge/br_netfilter.ko
> > > 
> > > 
> > > 
> > > But these modules are not included in my image.  For example, if I do
> > > "find / -name br_netfilter.ko" on my target install, nothing is
> > > returned.
> > > 
> > > 
> > > 
> > > My bblayers.conf is as follows:
> > > 
> > > 
> > > 
> > > === Start conf/bblayers.conf ===
> > > # LAYER_CONF_VERSION is increased each time build/conf/bblayers.conf #
> > > changes incompatibly LCONF_VERSION = "7"
> > > 
> > > 
> > > 
> > > BBPATH = "${TOPDIR}"
> > > BBFILES ?= ""
> > > 
> > > 
> > > 
> > > BBLAYERS ?= " \
> > > 
> > >   /home/lmuhlenkamp/oe-core/meta \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-python \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-gnome \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-filesystems \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-oe \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-networking \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-initramfs \
> > >   /home/lmuhlenkamp/meta-openembedded/meta-webserver \
> > >   /home/lmuhlenkamp/meta-intel \
> > >   /home/lmuhlenkamp/meta-virtualization \
> > >   /home/lmuhlenkamp/meta-cloud-services \
> > >   /home/lmuhlenkamp/meta-cloud-services/meta-openstack \
> > >   /home/lmuhlenkamp/meta-iot-cloud \
> > >   /home/lmuhlenkamp/meta-secure-core/meta-tpm \
> > >   /home/lmuhlenkamp/meta-stryker/common \
> > >   /home/lmuhlenkamp/meta-stryker/testing \
> > >   "
> > > 
> > > === End conf/bblayers.conf ===
> > > 
> > > 
> > > 
> > > The customizations to my local.conf file are as follows:
> > > 
> > > 
> > > 
> > > === Start local.conf excerpt ===
> > > MACHINE ?= "intel-corei7-64"
> > > IMAGE_FSTYPES += "live"
> > > NOISO = "0"
> > > IMAGE_INSTALL_append = " glibc-utils localedef"
> > > GLIBC_GENERATE_LOCALES = "el_GR.UTF-8 en_GB.UTF-8 en_US.UTF-8
> > > es_ES.UTF-8
> > > de_DE.UTF-8 fa_IR fr_FR.UTF-8 hr_HR.UTF-8 ja_JP.UTF-8 ja_JP.EUC-JP
> > > lt_LT.UTF-8 ru_RU.UTF-8 tr_TR.UTF-8"
> 
>  IMAGE_LINGUAS = "el-gr en-gb en-us
> 
> > > es-es de-de fa-ir fr-fr hr-hr ja-jp ja-jp.euc-jp lt-lt ru-ru tr-tr"
> > > DISTRO_FEATURES_append = " systemd virtualization"
> > > DISTRO_FEATURES_BACKFILL_CONSIDERED += "sysvinit"
> > > VIRTUAL-RUNTIME_init_manager = "systemd"
> > > VIRTUAL-RUNTIME_initscripts = "systemd-compat-units"
> > > DISTRO_FEATURES_append = " opengl"
> > > CORE_IMAGE_EXTRA_INSTALL += "rpm python3 python3-pip python3-flask
> > > python3-requests python3-coverage python3-pylint"
> > > CORE_IMAGE_EXTRA_INSTALL += "python-sphinx"
> > > CORE_IMAGE_EXTRA_INSTALL += "python-flake8"
> > > CORE_IMAGE_EXTRA_INSTALL += "python3-doxypypy"
> > > CORE_IMAGE_EXTRA_INSTALL += "trousers tpm-tools openssl-tpm-engine"
> > > KERNEL_FEATURES_append = " features/netfilter/netfilter.scc
> > > features/overlayfs/overlayfs.scc"
> 
>  KERNEL_ENABLE_CGROUPS = "1"
> 
> > > CORE_IMAGE_EXTRA_INSTALL += "docker"
> > > SERIAL_CONSOLES = "38400 tty1"
> > > CORE_IMAGE_EXTRA_INSTALL += "flaskhello"
> > > === End local.conf excerpt ===
> > > 
> > > 
> > > 
> > > I did not have the KERNEL_FEATURES_append line in when using the
> > > recipes-kernel/linux/linux-intel stuff.  I didn't have the
> > > recipes-kernel/linux/linux-intel directory available when using the
> > > KERNEL_FEATURES_append line in local.conf.  Based on what I read, they
> > > were mutually exclusive.
> > > 
> > > 
> > > 
> > > What am I missing?  Why are the kernel modules not being included in my
> > > image?
> > > 
> > > 
> > > 
> > > I did try tweaking my docker.cfg file in my custom kernel recipe to
> > > include cgroups, but that did not seem to change anything.  Here are
> > > the contents of my custom docker.cfg file
> > > 
> > > 
> > > 
> > > === Start
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/docker/docker.cfg
> > > === CONFIG_CGROUP_DEVICE=y
> > > 
> > > 
> > > 
> > > CONFIG_NETFILTER_XT_MATCH_ADDRTYPE=m
> > > CONFIG_IP_NF_FILTER=m
> > > CONFIG_NF_NAT=m
> > > CONFIG_NF_CONNTRACK_IPV4=y
> > > CONFIG_NF_CT_NETLINK=y
> > > 
> > > 
> > > 
> > > CONFIG_BRIDGE_NETFILTER=m
> > > CONFIG_XFRM_USER=m
> > > 
> > > 
> > > 
> > > CONFIG_DM_THIN_PROVISIONING=m
> > > 
> > > 
> > > 
> > > 
> > > CONFIG_IP_NF_NAT=m
> > > CONFIG_IP_NF_TARGET_MASQUERADE=m
> > > 
> > > 
> > > 
> > > CONFIG_OVERLAY_FS=y
> > > === End
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/docker/docker.cfg
> > > ===
> > > 
> > > 
> > > 
> > > And for completeness, here are the contents of the other 2 files in
> > > that recipe
> > > 
> > > 
> > > 
> > > === Start
> > > meta-stryker/common/recipes-kernel/linux/linux-intel_%.bbappend ===
> > > FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"
> 
>  SRC_URI +=
> 
> > > "file://kernel_baseline.scc"
> > > === End
> > > meta-stryker/common/recipes-kernel/linux/linux-intel_%.bbappend ===
> > > 
> > > 
> > > 
> > > === Start
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/kernel_baseline.scc
> > > === define KFEATURE_DESCRIPTION "Enable Features needed by docker in
> > > addition to LXC features"
> 
>  define KFEATURE_COMPATIBILITY board
> 
> > > kconf non-hardware docker.cfg
> > > === End
> > > meta-stryker/common/recipes-kernel/linux/linux-intel/kernel_baseline.s
> > > cc ===
> > > 
> > > 
> > > 
> > > Any help that would allow me to get docker functionality working in my
> > > openembedded linux image would be greatly appreciated.
> > > 
> > > 
> > > 
> > > Thank you
> > > 
> > > 
> > > 
> > > Lewis Muhlenkamp
> > > 
> > > 
> > > 
> > > --
> > > _______________________________________________
> > > Openembedded-devel mailing list
> > > Openembedded-devel at lists.openembedded.org
> > > http://lists.openembedded.org/mailman/listinfo/openembedded-devel

More information about the Openembedded-devel mailing list