openstack
/
sahara-image-elements
Code Issues Proposed changes

hadoop-cdh: simplify package installation 

Move the repository addition in a pre-install.d script, as that is the
right place for such operations (no more need to manual `apt-get update`
run), and switch it from curl to wget (already in use in that script);
as a result, wget is now installed in pre-install.d.

The above change makes the installation of hadoop-hdfs-namenode and
hadoop-hdfs-datanode possible using the declarative package-installs
way.

As a side effect of installing packages before setting up the hadoop
user and group, make sure to not create them if already present. The
packaging as of today takes care of creating the hadoop group, so guard
the setup in case it is not done anymore or will be done for the user
too.

Change-Id: I248872e77358f8388f4a66a54e53b2ffa727c7e3
This commit is contained in:
Pino Toscano 2015-05-18 15:42:57 +02:00
parent 113ccb320c
commit e1c6b32d1e
3 changed files with 38 additions and 40 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
elements/hadoop-cdh/package-installs.yaml
View File

@ -1,4 +1,4 @@
curl:
phase: post-install.d
wget:
phase: post-install.d
phase: pre-install.d
hadoop-hdfs-namenode:
hadoop-hdfs-datanode:

43
elements/hadoop-cdh/post-install.d/40-setup-hadoop
View File

@ -10,50 +10,19 @@ fi
set -eu
set -o pipefail
if [ "$DISTRO_NAME" != "ubuntu" ]; then
echo "Distro $DISTRO_NAME not supported by CDH. Exiting."
exit 1
fi
echo "Hadoop CDH setup begins for $DISTRO_NAME"
tmp_dir=/tmp/hadoop
echo "Creating hadoop user & group"
case "$DISTRO_NAME" in
ubuntu )
addgroup hadoop
adduser --ingroup hadoop --disabled-password --gecos GECOS hadoop
if ! getent group hadoop > /dev/null; then
addgroup hadoop
fi
if ! getent passwd hadoop > /dev/null; then
adduser --ingroup hadoop --disabled-password --gecos GECOS hadoop
fi
adduser hadoop sudo
;;
esac
echo "CDH 4 will be injected into image. Starting the download"
# Here more versions of CDH could be supported by downloading the right repository package.
wget -P $tmp_dir "http://archive.cloudera.com/cdh4/one-click-install/precise/amd64/cdh4-repository_1.0_all.deb"
if [ $? -ne 0 ]; then
echo -e "Could not find CDH 4.\nAborting"
exit 1
fi
# Pin packages from cloudera repository
cat >> /etc/apt/preferences.d/cloudera << EOF
Package: *
Pin: origin "archive.cloudera.com"
Pin-Priority: 800
EOF
case "$DISTRO_NAME" in
ubuntu )
dpkg -i $tmp_dir/cdh4-repository_1.0_all.deb
curl -s http://archive.cloudera.com/cdh4/ubuntu/precise/amd64/cdh/archive.key | sudo apt-key add -
sudo apt-get update
# Here the script could be expanded to install all CDH packages and not only HDFS.
install-packages hadoop-hdfs-namenode hadoop-hdfs-datanode
;;
esac
rm -r $tmp_dir
echo "Pre-configuring Hadoop"
cat >> /home/hadoop/.bashrc <<EOF

29
elements/hadoop-cdh/pre-install.d/40-cdh-repo Executable file
View File

@ -0,0 +1,29 @@
#!/bin/bash
if [ "${DIB_DEBUG_TRACE:-0}" -gt 0 ]; then
set -x
fi
set -eu
set -o pipefail
case "$DISTRO_NAME" in
ubuntu )
# Here more versions of CDH could be supported by downloading the right repository package.
wget -P /tmp "http://archive.cloudera.com/cdh4/one-click-install/precise/amd64/cdh4-repository_1.0_all.deb"
# Pin packages from cloudera repository
cat >> /etc/apt/preferences.d/cloudera << EOF
Package: *
Pin: origin "archive.cloudera.com"
Pin-Priority: 800
EOF
dpkg -i /tmp/cdh4-repository_1.0_all.deb
rm /tmp/cdh4-repository_1.0_all.deb
wget -O - http://archive.cloudera.com/cdh4/ubuntu/precise/amd64/cdh/archive.key | sudo apt-key add -
;;
*)
echo "Distro $DISTRO_NAME not supported by CDH. Exiting."
exit 1
;;
esac