Simple deployment refactoring - lots of changes

Removed remote SSH entirely from cluster deployment process
Trimmed boot.sh for CloudInit
Trimmed cm_boot_mysql.sh for CloudInit, added cluster deployment for testing here including relevant dependencies
Removed SSH, changed cluster deployment to local on CM host for deploy_on_oci.py.  Added a -S (SIMPLE) flag to deploy cluster without HA or Kerberos.

Terraform modified to support additional metadata fields

Author: smithzc

scripts/boot.sh

@@ -6,6 +6,7 @@ cdh_version=`curl -L http://169.254.169.254/opc/v1/instance/metadata/cdh_version
 cdh_major_version=`echo $cdh_version | cut -d '.' -f1`
 cm_version=`curl -L http://169.254.169.254/opc/v1/instance/metadata/cm_version`
 cm_major_version=`echo  $cm_version | cut -d '.' -f1`
+block_volume_count=`curl -L http://169.254.169.254/opc/v1/instance/metadata/block_volume_count`
 EXECNAME="TUNING"
 log "->TUNING START"
 sed -i.bak 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/selinux/config
@@ -16,8 +17,7 @@ yum install java-1.8.0-openjdk.x86_64 -y >> $LOG_FILE
 yum install krb5-workstation -y >> $LOG_FILE
 EXECNAME="KERBEROS"
 log "->krb5.conf"
-kdc_server='cdh-utility-1'
-kdc_fqdn=`host $kdc_server | gawk '{print $1}'`
+kdc_fqdn=${cm_fqdn}
 realm="hadoop.com"
 REALM="HADOOP.COM"
 log "-> CONFIG"
@@ -168,33 +168,16 @@ while [ "$detection_flag" = "0" ]; do
 			fi
 		fi
 	done;
-	log "->Sanity check, loop again"
-	sleep 30
-	sanity_detection_done="0"
-	sanity_volume_count="0"
-	for i in `seq 2 33`; do
-                if [ $sanity_detection_done = "0" ]; then
-                        iscsiadm -m discoverydb -D -t sendtargets -p 169.254.2.$i:3260 2>&1 2>/dev/null
-                        iscsi_chk=`echo -e $?`
-                        if [ $iscsi_chk = "0" ]; then
-                                siqn[${i}]=`iscsiadm -m discoverydb -D -t sendtargets -p 169.254.2.${i}:3260 | gawk '{print $2}'`
-                                continue
-                        else
-                                sanity_volume_count="${#siqn[@]}"
-                                log "--> Sanity Discovery Complete - ${#siqn[@]} volumes found"
-                                sanity_detection_done="1"
-                        fi
-                fi
-        done;
+	total_volume_count=$((block_volume_count+2))
 	if [ "$volume_count" = "0" ]; then 
-		log "-- $volume_count Block Volumes detected, sleeping 30 then retry --"
-		sleep 30
+		log "-- $volume_count Block Volumes detected, sleeping 15 then retry --"
+		sleep 15
 		continue
-	elif [ "$volume_count" != "$sanity_volume_count" ]; then
-		log "-- Sanity Check Failed - $sanity_volume_count Volumes found, $volume_count on first run.  Re-running --"
+	elif [ "$volume_count" != "$total_volume_count" ]; then
+		log "-- Sanity Check Failed - $volume_count Volumes found, $total_volume_count expected.  Re-running --"
 		sleep 15
 		continue
-	elif [ "$volume_count" = "$sanity_volume_count" ]; then 
+	elif [ "$volume_count" = "$total_volume_count" ]; then 
 		log "-- Setup for ${#iqn[@]} Block Volumes --"
 		for i in `seq 1 ${#iqn[@]}`; do
 			n=$((i+1))

scripts/cm_boot_mysql.sh

@@ -1,10 +1,15 @@
 #!/bin/bash
 LOG_FILE="/var/log/cloudera-OCI-initialize.log"
 log() { echo "$(date) [${EXECNAME}]: $*" >> "${LOG_FILE}" }
+cm_fqdn=`curl -L http://169.254.169.254/opc/v1/instance/metadata/cloudera_manager`
+cm_ip=`host ${cm_fqdn} | gawk '{print $4}'`
 cdh_version=`curl -L http://169.254.169.254/opc/v1/instance/metadata/cdh_version`
 cdh_major_version=`echo $cdh_version | cut -d '.' -f1`
 cm_version=`curl -L http://169.254.169.254/opc/v1/instance/metadata/cm_version`
 cm_major_version=`echo  $cm_version | cut -d '.' -f1`
+availability_domain=`curl -L http://169.254.169.254/opc/v1/instance/metadata/availability_domain`
+worker_shape=`curl -L http://169.254.169.254/opc/v1/instance/metadata/worker_shape`
+worker_disk_count=`curl -L http://169.254.169.254/opc/v1/instance/metadata/block_volume_count`
 EXECNAME="TUNING"
 log "-> START"
 sed -i.bak 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/selinux/config
@@ -36,8 +41,7 @@ log "-> INSTALL"
 yum -y install krb5-server krb5-libs krb5-workstation >> $LOG_FILE
 KERBEROS_PASSWORD="SOMEPASSWORD"
 SCM_USER_PASSWORD="somepassword"
-kdc_server=$(hostname)
-kdc_fqdn=`host $kdc_server | gawk '{print $1}'`
+kdc_server=${cm_fqdn}
 realm="hadoop.com"
 REALM="HADOOP.COM"
 log "-> CONFIG"
@@ -351,5 +355,35 @@ EXECNAME="Clouera Manager"
 log "->Starting Cloudera Manager"
 chown -R cloudera-scm:cloudera-scm /etc/cloudera-scm-server
 systemctl start cloudera-scm-server
+EXECNAME="Cluster Build"
+log "->Installing Python Pre-reqs"
+sudo yum install python python-pip -y >> $LOG_FILE
+sudo pip install --upgrade pip >> $LOG_FILE
+sudo pip install cm_client >> $LOG_FILE
+log "->Running Cluster Deployment"
+log "-->Host Discovery"
+detection_flag="0"
+w=1
+while [ $detection_flag = "0" ]; do
+	worker_lookup=`host cdh-worker-$w.private${availability_domain}.cdhvcn.oraclevcn.com`
+	worker_check=`echo -e $?`
+	if [ $worker_check = "0" ]; then 
+		worker_fqdn[$w]=`cdh-worker-$w.private${availability_domain}.cdhvcn.oraclevcn.com`
+		w=$((w+1))
+	else
+		detection_flag="1"
+	fi
+done;
+worker_fqdn_list=""
+num_workers=${#worker_fqdn[@]}
+for w in `seq 1 $num_workers`; do 
+	if [ $w = "1" ]; then 
+		worker_fqdn_list="${worker_fqdn[$w]}"
+	else
+		worker_fqdn_list=`echo "${worker_fqdn_list},${worker_fqdn[$w]}"`
+	fi
+done;
+echo "curl -L http://169.254.169.254/opc/v1/instance/metadata/extended_metadata | base64 -d " >> $LOG_FILE
+echo "python deploy_on_oci.py -B -S -m ${cm_ip} -i ${worker_fqdn_list} -d ${worker_disk_count} -w ${worker_shape} -n ${num_workers} -cdh ${cdh_version} -ad ${availability_domain}" >> $LOG_FILE
 EXECNAME="END"
 log "->DONE"