Metacat

#!/bin/bash

#

# Backup files needed for metacat.  This script creates a directory in /var/metacat/backup,

# backs up the postgres database, metacat data files, and certificate and web config files,

# then syncs those files to an Amazon S3 bucket for backup.

#

# To run this file, install it in /usr/sbin or a similar location and add an

# entry in the root user's crontab to run the command periodically.  The

# following crontab entry would run it every night at 2:00 am

# 0 2 * * *       /usr/sbin/backup-aws.sh >> $HOME/cron-output 2>&1

#

# This is really just an example script and may not work in your environment

# uless you modify it appropriately.

#

# 25 Apr 2017 Matt Jones

# Name of the metacat database in postgres to be backed up

DBNAME=metacat

# Number of days of backups to keep online, anything older is removed

DAYSTOKEEP=7

# AWS S3 bucket to be used for backup

BUCKET=s3://arcticdata.io/backup

# Location of the metacat.properties file

METACATPROPERTIESPATH=/var/lib/tomcat7/webapps/metacat/WEB-INF/metacat.properties

# Location of the apache configuration file

APACHECONF=/etc/apache2/sites-enabled

#Location of the server key

KEYLOCATION=/etc/ssl/private 

#Location of the server certificate

CERTLOCATION=/etc/ssl/certs/www_arcticdata_io.crt

#

# Below here lie demons

#

# Set up our umask to protect files from prying eyes

umask 007

# Make a temp dir for the backed up files

TAG=`date +%F-%H%M%S`

DATADIR="/var/metacat"

ARCHROOT="/var/metacat/metacat-backup"

mkdir -p $ARCHROOT

chgrp postgres $ARCHROOT

chmod g+rwxs $ARCHROOT

ARCHDIR="$ARCHROOT"

mkdir -p $ARCHDIR

# Locations of file lists already archived

DATALIST=${ARCHDIR}/data-list

METALIST=${ARCHDIR}/meta-list

# Shut down the tomcat server so nobody else changes anything while we backup

#/etc/init.d/tomcat7 stop

echo Generate a list of new metadata files since the last backup

if [ ! -e ${METALIST} ];

then

	aws s3 ls ${BUCKET}/documents/ |awk -F" " '{print $4}' > ${METALIST}

fi

diff --old-line-format="" --unchanged-line-format="" <(sort ${METALIST}) <(ls -1 ${DATADIR}/documents | sort) > ${METALIST}-new

echo Generate a list of new data files since the last backup

if [ ! -e ${DATALIST} ];

then

	aws s3 ls ${BUCKET}/data/ |awk -F" " '{print $4}' > ${DATALIST}

fi

diff --old-line-format="" --unchanged-line-format="" <(sort ${DATALIST}) <(ls -1 ${DATADIR}/data | sort) > ${DATALIST}-new

echo Copy the metacat.properties file to /var/metacat

cp $METACATPROPERTIESPATH $ARCHDIR

echo Backup postgres

su - postgres -c "pg_dumpall | gzip > $ARCHDIR/metacat-postgres-backup.gz"

echo Copy the apache configuration files

tar czhf $ARCHDIR/apache-config-backup.tgz $APACHECONF $KEYLOCATION $CERTLOCATION

echo Sync the backup directory to Amazon S3

echo Handle each of the subdirectories independently

aws s3 sync $DATADIR/certs $BUCKET/certs

aws s3 sync $DATADIR/dataone $BUCKET/dataone

aws s3 sync $DATADIR/inline-data $BUCKET/inline-data

aws s3 sync $DATADIR/logs $BUCKET/logs

aws s3 sync $DATADIR/.metacat $BUCKET/.metacat

aws s3 sync $DATADIR/metacat-backup $BUCKET/metacat-backup

#aws s3 sync $DATADIR/solr-home $BUCKET/solr-home

#aws s3 sync $DATADIR/tdb $BUCKET/tdb

#aws s3 sync $DATADIR/temporary $BUCKET/temporary

echo Backup metadata files to S3

cat ${METALIST}-new | xargs -n1 -P30 -I {} aws s3 cp ${DATADIR}/documents/{} $BUCKET/documents/{}

cat ${METALIST}-new >> ${METALIST}

echo Backup data files to S3

cat ${DATALIST}-new | xargs -n1 -P30 -I {} aws s3 cp ${DATADIR}/data/{} $BUCKET/data/{}

cat ${DATALIST}-new >> ${DATALIST}

# Restart tomcat

#/etc/init.d/tomcat7 start

# Clean up the temp files

#rm -rf $ARCHDIR

# clean up any of the backup files that are older than DAYSTOKEEP

#find $ARCHROOT -mtime +$DAYSTOKEEP -exec rm -f {} \;

echo "DONE backup for $TAG"