Merge ~ubuntu-docker-images/ubuntu-docker-images/+git/cassandra:initial-review into ~ubuntu-docker-images/ubuntu-docker-images/+git/cassandra:main

[Athos Ribeiro]

Addressing Sergio's comments from the mailing list review at https://lists.launchpad.net/ubuntu-docker-images/msg00031.html

Two points are not being addressed here

1) Moving all RUN commands from the final build stage into a single RUN command:
This requires changes in the snap (marking the binary file as executable).

2) matching any Java version when setting the JAVA environment variable:
I'd rather explicitly tell what version to use in case the image ends up containing multiple Java versions installed by whatever reason (if this is ever the case, it should be fixed, but we may miss it and and up shipping a container that runs cassandra on an unexpected Java version. WDYT?)

[Athos Ribeiro (athos-ribeiro)]

I forgot to also update the description, as requested in the review (that should be done now).

Moreover, on the CQLVERSION matter: that seems to be specific to the container image used for the sqlsh CLI right?

In their entrypoint, I see they set "export CQLVERSION=${CQLVERSION:-"3.4.4"}".

And in Cassandra cqlsh.py, we have "by default the highest version supported by the server will be used" for the --cqlversion description.

I'd expect the user to use a client compatible with the version of the server being deployed.

I also intend to snap the cqlsh in the future (ship it within the cassandra snap).

[Sergio Durigan Junior (sergiodj)]

Thanks for the MP, Athos!

It looks great. I'm leaving a few comments about indentation, but otherwise I'm happy with the changes you've made. As for the comments you posted:

1) ACK. I will review the snap change when you put it up, and then we can join the RUN statements.

2) Agreed :-).

3) About the CQLVERSION, I agree that it's specific to the cqlsh image. I was just frustrated that I needed to do some digging in order to figure out why the "docker run" wasn't working. But we can revisit this later if needed.

review: Needs Fixing

[Athos Ribeiro (athos-ribeiro)]

I merged the RUN, VOLUME, and ENV commands to reduce the number of layers generated here.

[Sergio Durigan Junior (sergiodj)]

Thanks, Athos. This looks better now :-). There's still the indentation nit to fix, but I think the Dockerfile is great otherwise.

[Athos Ribeiro (athos-ribeiro)]

Done!

Thanks, Sergio :)

[Sergio Durigan Junior (sergiodj)]

Thanks, Athos.

This LGTM now (modulo the "curl" command which will obviously need adjustment when you upload the snap).

review: Approve

diff --git a/Dockerfile b/Dockerfile
index d1cf855..ef4d6da 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -3,35 +3,31 @@
 # cassandra.yaml logback.xml cassandra-env.sh jvm.options
 FROM ubuntu:focal AS snap-installer
 
-RUN apt-get update && \
-    DEBIAN_FRONTEND=noninteractive apt-get dist-upgrade -y && \
-    DEBIAN_FRONTEND=noninteractive apt-get install -y \
-    curl ca-certificates squashfs-tools
-
-# taken from https://snapcraft.io/docs/build-on-docker
-# Alternatively, we can install snapd, and issue `snap download cassandra`
-RUN curl -L 'https://launchpad.net/~athos-ribeiro/+snap/cassandra-test/+build/1430794/+files/cassandra_4.0-rc1_amd64.snap' --output cassandra.snap
-
-RUN mkdir -p /snap && unsquashfs -d /snap/cassandra cassandra.snap
+RUN set -eux; \
+	apt-get update && \
+	DEBIAN_FRONTEND=noninteractive apt-get full-upgrade -y && \
+	DEBIAN_FRONTEND=noninteractive apt-get install -y \
+		curl ca-certificates squashfs-tools && \
+	# taken from https://snapcraft.io/docs/build-on-docker
+	# Alternatively, we can install snapd, and issue `snap download cassandra`
+	curl -L 'https://launchpad.net/~athos-ribeiro/+snap/cassandra-test/+build/1437307/+files/cassandra_4.0-rc1_amd64.snap' --output cassandra.snap && \
+	mkdir -p /snap && unsquashfs -d /snap/cassandra cassandra.snap
 
 
 FROM ubuntu:focal
 
 LABEL maintainer="Ubuntu Server team <[e-mail]>"
 
-ENV TZ UTC
-
-ENV CASSANDRA_HOME /usr/share/cassandra
-ENV STATIC_CONF /etc/cassandra
-ENV CASSANDRA_CONF /etc/cassandra
-ENV CASSANDRA_LOG_DIR /var/log/cassandra
-ENV CASSANDRA_DATA /var/lib/cassandra
-ENV CLASSPATH $CASSANDRA_CONF
-ENV CASSANDRA_INCLUDE ""
-ENV JAVA_VERSION 8
-ENV cassandra_storagedir "$CASSANDRA_DATA"
-
-RUN mkdir -p $CASSANDRA_CONF $CASSANDRA_LOG_DIR $CASSANDRA_DATA
+ENV TZ=UTC \
+	CASSANDRA_HOME=/usr/share/cassandra \
+	STATIC_CONF=/etc/cassandra \
+	CASSANDRA_CONF=/etc/cassandra \
+	CASSANDRA_LOG_DIR=/var/log/cassandra \
+	CASSANDRA_DATA=/var/lib/cassandra \
+	CASSANDRA_INCLUDE="" \
+	JAVA_VERSION=8
+ENV CLASSPATH=$CASSANDRA_CONF \
+	cassandra_storagedir="$CASSANDRA_DATA"
 
 COPY --from=snap-installer /snap/cassandra/usr/sbin/cassandra /usr/sbin/cassandra
 COPY --from=snap-installer /snap/cassandra/usr/share/cassandra /usr/share/cassandra
@@ -45,32 +41,32 @@ RUN set -eux; \
 	apt-get update; \
 	DEBIAN_FRONTEND=noninteractive apt-get full-upgrade -y; \
 	DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
-  openjdk-8-jre-headless ca-certificates tzdata \
-# solves warning: "jemalloc shared library could not be preloaded to speed up memory allocations"
+		openjdk-8-jre-headless ca-certificates tzdata \
+		# solves warning: "jemalloc shared library could not be preloaded to speed up memory allocations"
 		libjemalloc2 \
-# "free" is used by cassandra-env.sh
+		# "free" is used by cassandra-env.sh
 		procps \
-# "ip" is not required by Cassandra itself, but is commonly used in scripting Cassandra's configuration (since it is so fixated on explicit IP addresses)
+		# "ip" is not required by Cassandra itself, but is commonly used in scripting Cassandra's configuration (since it is so fixated on explicit IP addresses)
 		iproute2 \
-# Cassandra will automatically use numactl if available
-#   https://github.com/apache/cassandra/blob/18bcda2d4c2eba7370a0b21f33eed37cb730bbb3/bin/cassandra#L90-L100
-#   https://github.com/apache/cassandra/commit/604c0e87dc67fa65f6904ef9a98a029c9f2f865a
+		# Cassandra will automatically use numactl if available
+		#   https://github.com/apache/cassandra/blob/18bcda2d4c2eba7370a0b21f33eed37cb730bbb3/bin/cassandra#L90-L100
+		#   https://github.com/apache/cassandra/commit/604c0e87dc67fa65f6904ef9a98a029c9f2f865a
 		numactl \
-	; \
+		; \
 	DEBIAN_FRONTEND=noninteractive apt-get remove --purge --auto-remove -y; \
 	rm -rf /var/lib/apt/lists/*; \
-# https://issues.apache.org/jira/browse/CASSANDRA-15767 ("bin/cassandra" only looks for "libjemalloc.so" or "libjemalloc.so.1" which doesn't match our "libjemalloc.so.2")
+	# https://issues.apache.org/jira/browse/CASSANDRA-15767 ("bin/cassandra" only looks for "libjemalloc.so" or "libjemalloc.so.1" which doesn't match our "libjemalloc.so.2")
 	libjemalloc="$(readlink -e /usr/lib/*/libjemalloc.so.2)"; \
 	ln -sT "$libjemalloc" /usr/local/lib/libjemalloc.so; \
 	ldconfig; \
-  chmod +x /usr/sbin/cassandra
+	mkdir -p $CASSANDRA_CONF $CASSANDRA_LOG_DIR $CASSANDRA_DATA
 
 ENV JVM_OPTS "$JVM_OPTS -Dcassandra.config=file://$CASSANDRA_CONF/cassandra.yaml"
 
 COPY entrypoint.sh /usr/local/bin/
 
-VOLUME /var/lib/cassandra
-VOLUME /var/log/cassandra
+VOLUME /var/lib/cassandra \
+	/var/log/cassandra
 
 ENTRYPOINT ["entrypoint.sh"]
 # 7000: intra-node communication
diff --git a/data/cassandra.yaml b/data/cassandra.yaml
index 281a5cc..92a0d7b 100644
--- a/data/cassandra.yaml
+++ b/data/cassandra.yaml
@@ -2,10 +2,11 @@ application: Cassandra
 main: true
 repo: cassandra
 description: >
-  Apache Cassandra is a highly-scalable partitioned row store. Rows are organized into tables with a required primary key.
-  [Partitioning](https://cwiki.apache.org/confluence/display/CASSANDRA2/Partitioners) means that Cassandra can distribute your data across multiple machines in an application-transparent matter. Cassandra will automatically repartition as machines are added and removed from the cluster.
-  [Row store](https://cwiki.apache.org/confluence/display/CASSANDRA2/DataModel) means that like relational databases, Cassandra organizes data by rows and columns. The Cassandra Query Language (CQL) is a close relative of SQL.
-  For more information, see the [Apache Cassandra web site](http://cassandra.apache.org/).
+  Apache Cassandra is an open source NoSQL distributed database trusted by
+  thousands of companies for scalability and high availability without
+  compromising performance. Linear scalability and proven fault-tolerance on
+  commodity hardware or cloud infrastructure make it a good platform for
+  mission-critical data.
 version: "4.0-rc1"
 base: Ubuntu 20.04
 architectures:
diff --git a/entrypoint.sh b/entrypoint.sh
index 5cc4c47..59c947f 100755
--- a/entrypoint.sh
+++ b/entrypoint.sh
@@ -1,29 +1,12 @@
 #!/bin/bash
 
-for jar in $SNAP/usr/share/cassandra/lib/*.jar; do
-    CLASSPATH="$CLASSPATH:$jar"
+for jar in /usr/share/cassandra/lib/*.jar; do
+	CLASSPATH="$CLASSPATH:$jar"
 done
 
 export CLASSPATH
 
-case $(arch) in
-    x86_64)
-        export ARCH=amd64
-        ;;
-    aarch64)
-        export ARCH=arm64
-        ;;
-    ppc64le)
-        export ARCH=ppc64el
-        ;;
-    s390x)
-        export ARCH=s390x
-        ;;
-    *)
-        echo "Unsupported architecture $(arch)"
-        exit 1
-        ;;
-esac
+export ARCH=$(dpkg --print-architecture)
 export JAVA=/usr/lib/jvm/java-8-openjdk-"$ARCH"/jre/bin/java
 
 # From upstream entrypoint