Minimize docker image

docker/dev/find_fastest_resources.sh

@@ -1,4 +1,5 @@
 #!/bin/bash
+
 # Copyright 2020 The SQLFlow Authors. All rights reserved.
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -36,7 +37,7 @@ function get_domain_from_url() {
 
 # Find the fastest URL. The parameter consists of URLS separated by whitespace.
 function find_fastest_url() {
-    local speed=99999.9
+    local speed=99999
     # shellcheck disable=SC2068
     for i in $@; do
         local domain
@@ -48,10 +49,11 @@ function find_fastest_url() {
         cur_speed=$(ping -c 4 -W 2 "$domain" | tail -1 \
                            | grep "/avg/" | awk '{print $4}'\
                            | cut -d '/' -f 2)
-        cur_speed=${cur_speed:-99999.9}
+        cur_speed=${cur_speed:-99999}
+        cur_speed=${cur_speed/.*/}
 
         # c.f. https://stackoverflow.com/a/31087503/724872
-        if (( $(echo "$cur_speed < $speed" | bc -l) )); then
+        if [[ $cur_speed -lt $speed ]]; then
             local best_domain="$i"
             speed="$cur_speed"
         fi
@@ -245,3 +247,17 @@ function choose_fastest_pip_source() {
     find_fastest_pip_mirror > "$HOME"/.pip/pip.conf
 }
 
+function choose_fastest_alpine_source() {
+    default="http://dl-cdn.alpinelinux.org"
+    read -r -d '\t' urls <<EOM
+$default
+http://mirrors.tuna.tsinghua.edu.cn
+\t
+EOM
+
+    best=$(find_fastest_url $urls)
+    if [[ "$best" != "$default" ]]; then
+        sed -i "s=http://dl-cdn.alpinelinux.org=$best=g" /etc/apk/repositories
+    fi
+}
+

docker/mysql/Dockerfile

@@ -21,12 +21,7 @@
 # populating datasets, and the container running sqlflow:ci can use
 # inotifywait to wait for the creation of this file before it go on
 # running tests.
-FROM ubuntu:18.04
-
-COPY docker/dev/find_fastest_resources.sh /usr/local/bin/
-
-COPY docker/mysql/install-mysql-server.bash /
-RUN /install-mysql-server.bash
+FROM alpine:3.12
 
 # Install sample datasets for CI and demo.
 COPY doc/datasets/popularize_churn.sql \
@@ -37,6 +32,12 @@ COPY doc/datasets/popularize_churn.sql \
      doc/datasets/create_model_db.sql \
      /datasets/
 
+COPY docker/dev/find_fastest_resources.sh /usr/local/bin/
+RUN /bin/sh -c "source find_fastest_resources.sh && \
+    choose_fastest_alpine_source"
+
+RUN apk add --no-cache mysql mysql-client >/dev/null
+
 VOLUME /var/lib/mysql
 
 ARG MYSQL_PORT="3306"

docker/mysql/start.bash

@@ -1,4 +1,5 @@
-#!/bin/bash
+#!/bin/sh
+
 # Copyright 2020 The SQLFlow Authors. All rights reserved.
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -14,14 +15,38 @@
 
 set -e
 
+
+echo "Init mysqld if needed ..."
+if [[ -d "/docker-entrypoint-initdb.d" ]]; then
+    echo "Skip"
+else
+    mkdir -p /var/run/mysqld
+    mkdir -p /var/lib/mysql
+    chown mysql:mysql /var/run/mysqld
+    chown mysql:mysql /var/lib/mysql
+    mkdir -p /docker-entrypoint-initdb.d
+
+    mysql_install_db --user=mysql --datadir=/var/lib/mysql >dev/null
+    mysqld --user=mysql --bootstrap --verbose=0 \
+        --skip-name-resolve --skip-networking=0 >/dev/null <<EOF
+FLUSH PRIVILEGES;
+DELETE FROM mysql.user;
+GRANT ALL ON *.* TO 'root'@'%' identified by 'root' WITH GRANT OPTION;
+DROP DATABASE IF EXISTS test;
+FLUSH PRIVILEGES;
+EOF
+fi
+
+
 echo "Start mysqld ..."
 # Important to make mysqld bind to 0.0.0.0 -- all IPs.  I explained
 # the reason in https://stackoverflow.com/a/61887788/724872.
 MYSQL_HOST=${MYSQL_HOST:-0.0.0.0}
 MYSQL_PORT=${MYSQL_PORT:-3306}
-sed -i "s/.*bind-address.*/bind-address = $MYSQL_HOST/" \
-    /etc/mysql/mysql.conf.d/mysqld.cnf
-service mysql start
+
+nohup mysqld --user=mysql --console \
+    --skip-name-resolve --skip-networking=0 >/dev/null 2>&1 &
+sleep 2
 
 
 echo "Sleep until MySQL server is ready ..."
@@ -31,26 +56,16 @@ until mysql -u root -proot \
             --port "$MYSQL_PORT" \
             -e ";" ; do
     sleep 1
-    read -r -p "Can't connect, retrying..."
+    echo "Can't connect, retrying..."
 done
 
 
-# Grant all privileges to all the remote hosts so that the sqlflow
-# server can be scaled to more than one replicas.
-#
-# NOTE: should notice this authorization on the production
-# environment, it's not safe.
-mysql -uroot -proot \
-      -e "GRANT ALL PRIVILEGES ON *.* TO 'root'@'' IDENTIFIED BY 'root' WITH GRANT OPTION;"
-
-
-# FIXME(typhoonzero): should let docker-entrypoint.sh do this work
+echo "Populate datasets ..."
 for f in /datasets/*; do
-    echo "Populate datasets $f ..."
-    mysql -uroot -proot \
-          --host "$MYSQL_HOST" --port "$MYSQL_PORT" \
-          < "$f"
+    echo "$f"
+    mysql -uroot -proot < "$f"
 done
+echo "Done."
 
 
 # If we run the contaienr with -v host_dir:/work, then the following
@@ -59,4 +74,10 @@ done
 # file using the trick https://unix.stackexchange.com/a/185370/325629.
 mkdir -p /work && touch /work/mysql-inited
 
-sleep infinity
+
+# c.f. https://stackoverflow.com/questions/2935183/bash-infinite-sleep-infinite-blocking for BusyBox
+echo "Serving ..."
+while true; 
+    do sleep 1d;
+done
+

pkg/workflow/argo/fetch_test.go

@@ -215,9 +215,9 @@ func TestFetch(t *testing.T) {
 	a.Contains(concatedLogs, "SQLFlow Step: [3/3] Status: Succeeded")
 	// confirm columns and rows of sql: SELECT 1;
 	a.Equal([]string{"1"}, columns)
-	v := &wrappers.Int64Value{}
+	v := &wrappers.Int32Value{}
 	a.NoError(ptypes.UnmarshalAny(rows[0][0], v))
-	a.Equal(v.GetValue(), int64(1))
+	a.Equal(v.GetValue(), int32(1))
 }
 
 func waitUntilPodRunning(podID string) error {