nfs-common install issue

[andyzhangx]

What happened:

blob-csi-driver/pkg/blobplugin/Dockerfile

Lines 28 to 29 in 06df5fe

	# this is a workaround to install nfs-common and don't quit with error
	RUN apt update && apt install mount nfs-common -y || true

#16 9.211 Setting up nfs-common (1:1.3.4-2.5+deb10u1) ...
#16 9.233 /var/lib/dpkg/info/nfs-common.postinst: 7: /var/lib/dpkg/info/nfs-common.postinst: ucf: not found
#16 9.233 dpkg: error processing package nfs-common (--configure):
#16 9.233  installed nfs-common package post-installation script subprocess returned error exit status 127
#16 9.234 Processing triggers for libc-bin (2.28-10) ...
#16 9.266 Errors were encountered while processing:
#16 9.266  nfs-common
#16 9.302 E: Sub-process /usr/bin/dpkg returned an error code (1)
#16 ERROR: executor failed running [/bin/sh -c apt update && apt install mount nfs-common -y]: buildkit-runc did not terminate successfully
------
 > [stage-1 6/6] RUN apt update && apt install mount nfs-common -y:
------
failed to solve: rpc error: code = Unknown desc = executor failed running [/bin/sh -c apt update && apt install mount nfs-common -y]: buildkit-runc did not terminate successfully

What you expected to happen:

How to reproduce it:

Anything else we need to know?:

Environment:

• CSI Driver version:
• Kubernetes version (use kubectl version):
• OS (e.g. from /etc/os-release):
• Kernel (e.g. uname -a):
• Install tools:
• Others:

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[andyzhangx]

cc @dr0pdb

[dr0pdb]

/assign

[dr0pdb]

@andyzhangx I ran the Dockerfile locally and didn't get any error.

┌[satiwa☮MININT-UMBS8MU]-(~/.gvm/pkgsets/go1.15/global/src/sigs.k8s.io/blob-csi-driver/pkg/blobplugin)-[git://nfs-common-issue ✗]-
└> cat blob.Dockerfile
FROM us.gcr.io/k8s-artifacts-prod/build-image/debian-base-amd64:v2.1.0
RUN apt-get update && clean-install ca-certificates pkg-config libfuse-dev cmake libcurl4-gnutls-dev libgnutls28-dev uuid-dev libgcrypt20-dev wget
RUN wget -O /tmp/packages-microsoft-prod.deb https://packages.microsoft.com/config/ubuntu/16.04/packages-microsoft-prod.deb && dpkg -i /tmp/packages-microsoft-prod.deb && apt-get update && apt install blobfuse fuse -y && rm -f /tmp/packages-microsoft-prod.deb
RUN apt remove wget -y
# this is a workaround to install nfs-common & nfs-kernel-server and don't quit with error
RUN apt update && apt install nfs-common nfs-kernel-server -y || true
LABEL maintainers="andyzhangx"
LABEL description="Azure Blob Storage CSI driver"%
┌[satiwa☮MININT-UMBS8MU]-(~/.gvm/pkgsets/go1.15/global/src/sigs.k8s.io/blob-csi-driver/pkg/blobplugin)-[git://nfs-common-issue ✗]-
└> docker build --no-cache -t andyzhangx/blobnfs:v1 -f ./blob.Dockerfile .
[+] Building 110.4s (9/9) FINISHED
 => [internal] load build definition from blob.Dockerfile                                        0.1s
 => => transferring dockerfile: 43B                                                              0.0s
 => [internal] load .dockerignore                                                                0.1s
 => => transferring context: 2B                                                                  0.0s
 => [internal] load metadata for us.gcr.io/k8s-artifacts-prod/build-image/debian-base-amd64:v2.  1.0s
 => CACHED [1/5] FROM us.gcr.io/k8s-artifacts-prod/build-image/debian-base-amd64:v2.1.0@sha256:  0.0s
 => [2/5] RUN apt-get update && clean-install ca-certificates pkg-config libfuse-dev cmake lib  67.5s
 => [3/5] RUN wget -O /tmp/packages-microsoft-prod.deb https://packages.microsoft.com/config/u  14.1s
 => [4/5] RUN apt remove wget -y                                                                 2.1s
 => [5/5] RUN apt update && apt install nfs-common nfs-kernel-server -y || true                 23.2s
 => exporting to image                                                                           2.2s
 => => exporting layers                                                                          2.1s
 => => writing image sha256:85a9d729ff0dd1075a6209f4e2f667a4817b62a2d5a9d6dd2a65f6d074e0f047     0.0s
 => => naming to docker.io/andyzhangx/blobnfs:v1

[andyzhangx]

can you change RUN apt update && apt install nfs-common nfs-kernel-server -y || true to RUN apt update && apt install nfs-common nfs-kernel-server -y and try again?

[dr0pdb]

Yeah, I was able to reproduce. I'll try to fix.

...
#8 28.33 /var/lib/dpkg/info/nfs-common.postinst: 7: /var/lib/dpkg/info/nfs-common.postinst: ucf: not found
#8 28.33 dpkg: error processing package nfs-common (--configure):
#8 28.33  installed nfs-common package post-installation script subprocess returned error exit status 127
#8 28.33 dpkg: dependency problems prevent configuration of nfs-kernel-server:
#8 28.33  nfs-kernel-server depends on nfs-common (= 1:1.3.4-2.5+deb10u1); however:
#8 28.33   Package nfs-common is not configured yet.
#8 28.33
#8 28.33 dpkg: error processing package nfs-kernel-server (--configure):
#8 28.33  dependency problems - leaving unconfigured
#8 28.33 Processing triggers for libc-bin (2.28-10) ...
#8 28.45 Errors were encountered while processing:
#8 28.45  nfs-common
#8 28.45  nfs-kernel-server
#8 28.47 E: Sub-process /usr/bin/dpkg returned an error code (1)
------
executor failed running [/bin/sh -c apt update && apt install nfs-common nfs-kernel-server -y]: exit code: 100

[sozercan]

I was investigating why this is happening, looks like /usr/bin/ucf is a script that has a /bin/bash shebang. However, debian-slim based image does not contain bash which is why it's failing.

This is already fixed upstream in debian sid - https://metadata.ftp-master.debian.org/changelogs//main/u/ucf/ucf_3.0043_changelog

  * Bug fix: "ucf has /bin/bash shebangs but does not depend on bash",
    thanks to James Le Cuirot (Closes: #952448). It actually does not need
    bash, so this was easy to do.

But fix is not available in buster package yet.
https://metadata.ftp-master.debian.org/changelogs//main/u/ucf/ucf_3.0038+nmu1_changelog

[andyzhangx]

not fixed

[k8s-triage-robot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale

[sozercan]

/remove-lifecycle stale

[jchauncey]

any update here?

[andyzhangx]

any update here?

is this a blocking issue?

[jchauncey]

Ill DM you on teams about it.

[andyzhangx]

azure file Dockerfile is already using debian-base:bullseye-v1.0.0, so maybe it's already fixed?
https://github.com/kubernetes-sigs/azurefile-csi-driver/blob/d0bcaf51c17f32672f7135972d8860b5f57403d1/pkg/azurefileplugin/Dockerfile#L15-L21

FROM k8s.gcr.io/build-image/debian-base:bullseye-v1.0.0

ARG ARCH
COPY ./_output/${ARCH}/azurefileplugin /azurefileplugin

RUN apt update && apt-mark unhold libcap2
RUN clean-install ca-certificates cifs-utils util-linux e2fsprogs mount udev xfsprogs nfs-common

[jchauncey]

But that wont get rolled out until the next release right? Doesn't help existing images.