HDDS-14230. Create slim ozone image #47

rich7420 · 2025-12-21T12:00:22Z

To trim Ozone docker image size

ci test:
https://github.com/apache/ozone-docker/actions/runs/20455995904/job/58778071961

adoroszlai · 2025-12-22T15:48:56Z

@rich7420 can you please share image size before/after this change? (Like I mentioned in apache/ozone#9538, deleting files in a separate RUN step does not reduce image size.)

rich7420 · 2025-12-23T02:25:11Z

@adoroszlai , my mistake , I've added a slim dockerfile.
Image │ Logical Size │ Compressed Size
Original │ 2.34GB │ 0.82 GB
Slim │ 1.87GB │ 0.61 GB

adoroszlai

Thanks @rich7420 for the patch.

We'll need changes in both docker-ozone-runner and docker-ozone, so I've created subtasks for HDDS-13426. This change would then belong to HDDS-14230, and can be done after HDDS-14229.

It should target the slim branch, editing Dockerfile, so that workflows for CI check and publishing the image to container registries automatically pick it up.

Also, please enable GitHub Actions workflows in your forks of both repos.

adoroszlai · 2025-12-23T08:07:10Z

Dockerfile.slim

+ARG OZONE_VERSION=2.0.0
+ARG OZONE_URL="https://www.apache.org/dyn/closer.lua?action=download&filename=ozone/${OZONE_VERSION}/ozone-${OZONE_VERSION}.tar.gz"
+
+FROM alpine:latest AS extractor


We don't need the extractor stage, deletion can be added to the existing RUN curl, tar command.

adoroszlai · 2025-12-23T08:07:18Z

Dockerfile.slim

+    curl -LSs -o ozone.tar.gz "$OZONE_URL" && \
+    tar zxf ozone.tar.gz && \
+    cd ozone-${OZONE_VERSION} && \
+    rm -rf docs examples share/doc share/man licenses LICENSE.txt NOTICE.txt README.md HISTORY.md SECURITY.md CONTRIBUTING.md compose kubernetes share/ozone/byteman && \


Please list dirs/files to be deleted one per line, in alphabetical order. This makes it easier to understand and reduces diff when making changes later.

I think we should keep licenses LICENSE.txt NOTICE.txt.

adoroszlai · 2025-12-23T08:08:51Z

Dockerfile.slim

+    tar zxf ozone.tar.gz && \
+    cd ozone-${OZONE_VERSION} && \
+    rm -rf docs examples share/doc share/man licenses LICENSE.txt NOTICE.txt README.md HISTORY.md SECURITY.md CONTRIBUTING.md compose kubernetes share/ozone/byteman && \
+    find . -type f \( -name "ozone-filesystem-hadoop2-*.jar" -o -name "ozone-filesystem-hadoop3-*.jar" -o -name "ozone-filesystem-hadoop3-client-*.jar" -o -name "*-tests.jar" -o -name "*-test.jar" -o -name "*test*.jar" -o -name "*-docs-*.jar" -o -name "*-shaded.jar" -o -name "*-all.jar" -o -name "*-fat.jar" \) -delete && \


List of ozone-filesystem-hadoop2 etc. can be simplified to ozone-filesystem-hadoop*.jar.

"*test*.jar" covers the other two tests conditions.

oh you're right!

Also, "*-shaded.jar", "*-all.jar" and "*-fat.jar" are unnecessary.

adoroszlai · 2025-12-23T08:12:35Z

Dockerfile.slim

+    cd ozone-${OZONE_VERSION} && \
+    rm -rf docs examples share/doc share/man licenses LICENSE.txt NOTICE.txt README.md HISTORY.md SECURITY.md CONTRIBUTING.md compose kubernetes share/ozone/byteman && \
+    find . -type f \( -name "ozone-filesystem-hadoop2-*.jar" -o -name "ozone-filesystem-hadoop3-*.jar" -o -name "ozone-filesystem-hadoop3-client-*.jar" -o -name "*-tests.jar" -o -name "*-test.jar" -o -name "*test*.jar" -o -name "*-docs-*.jar" -o -name "*-shaded.jar" -o -name "*-all.jar" -o -name "*-fat.jar" \) -delete && \
+    find . -type d \( -name "test*" -o -name "tests" -o -name "*test" -o -name "__pycache__" \) -exec rm -rf {} + 2>/dev/null || true && \


Delete tests and smoketest in the first rm -rf command above. __pycache__ does not need to be deleted.

adoroszlai · 2025-12-23T08:14:30Z

Dockerfile.slim

+    rm -rf docs examples share/doc share/man licenses LICENSE.txt NOTICE.txt README.md HISTORY.md SECURITY.md CONTRIBUTING.md compose kubernetes share/ozone/byteman && \
+    find . -type f \( -name "ozone-filesystem-hadoop2-*.jar" -o -name "ozone-filesystem-hadoop3-*.jar" -o -name "ozone-filesystem-hadoop3-client-*.jar" -o -name "*-tests.jar" -o -name "*-test.jar" -o -name "*test*.jar" -o -name "*-docs-*.jar" -o -name "*-shaded.jar" -o -name "*-all.jar" -o -name "*-fat.jar" \) -delete && \
+    find . -type d \( -name "test*" -o -name "tests" -o -name "*test" -o -name "__pycache__" \) -exec rm -rf {} + 2>/dev/null || true && \
+    find . -type f \( -name "*.md" -o -name "*.txt" -o -name "*.pyc" \) ! -path "*/etc/*" ! -path "*/bin/*" ! -path "*/sbin/*" ! -path "*/libexec/*" -delete 2>/dev/null || true


This command is not needed.

rich7420 · 2025-12-23T08:43:17Z

@adoroszlai thanks for the support!

rich7420 · 2025-12-23T08:51:06Z

Here's the latest result.
Size: 1.88GB
Compressed: 0.62 GB

adoroszlai · 2025-12-25T09:10:01Z

Dockerfile.slim

+# limitations under the License.
+
+ARG OZONE_RUNNER_IMAGE=apache/ozone-runner
+ARG OZONE_RUNNER_VERSION=20250410-1-jdk21


Now this can be set to 20251225-1-jdk21-slim to use the smaller base image.

adoroszlai · 2025-12-25T09:17:50Z

Dockerfile.slim

+    sudo find . -type f \( \
+        -name "*-all.jar" -o \
+        -name "*-docs-*.jar" -o \
+        -name "*-fat.jar" -o \
+        -name "*-shaded.jar" -o \
+        -name "*test*.jar" -o \
+        -name "ozone-filesystem-hadoop*.jar" \
+    \) -delete


all, fat, shaded and test are unnecessary.

We can replace find with plain delete, because jars are in a single directory.

share/ozone/lib/*-docs-*.jar \ share/ozone/lib/ozone-filesystem-hadoop*.jar \

adoroszlai · 2025-12-25T09:22:46Z

Dockerfile.slim

+        share/doc \
+        share/man \


I don't see these in the tarball, are they necessary?

yap, those are unnecessary

Dockerfile

Trim Ozone docker image size

fe014ec

update

e3c276a

adoroszlai reviewed Dec 23, 2025

View reviewed changes

adoroszlai changed the base branch from latest to slim December 23, 2025 08:30

adoroszlai changed the title ~~HDDS-13426. Trim Ozone docker image size~~ HDDS-14230. Create slim ozone image Dec 23, 2025

adoroszlai marked this pull request as draft December 23, 2025 08:30

follow the suggestions

ea300aa

rich7420 marked this pull request as ready for review December 23, 2025 08:56

rich7420 marked this pull request as draft December 23, 2025 08:57

adoroszlai reviewed Dec 25, 2025

View reviewed changes

rich7420 added 2 commits December 25, 2025 17:31

update

c50c9f2

remove redundant part

cad10a9

adoroszlai reviewed Dec 26, 2025

View reviewed changes

Dockerfile Outdated Show resolved Hide resolved

Dockerfile Outdated Show resolved Hide resolved

Needs python, combine rm commands

a939c1b

adoroszlai reviewed Dec 26, 2025

View reviewed changes

Dockerfile Outdated Show resolved Hide resolved

Needs ncat

1ecc39f

adoroszlai marked this pull request as ready for review December 26, 2025 08:09

adoroszlai approved these changes Dec 26, 2025

View reviewed changes

adoroszlai merged commit 1ad65f1 into apache:slim Dec 26, 2025
2 checks passed

HDDS-14230. Create slim ozone image #47

HDDS-14230. Create slim ozone image #47

Uh oh!

Conversation

rich7420 commented Dec 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

adoroszlai commented Dec 22, 2025

Uh oh!

rich7420 commented Dec 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

adoroszlai left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rich7420 commented Dec 23, 2025

Uh oh!

rich7420 commented Dec 23, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

rich7420 commented Dec 21, 2025 •

edited

Loading

rich7420 commented Dec 23, 2025 •

edited

Loading