1. High Availability Overview and Best Practices
  2. MAA Platinum and Oracle GoldenGate Best Practices
  3. Cloud MAA Platinum: Oracle GoldenGate Microservices Architecture Integrated with Active Data Guard
  4. Task 3 - Configure Oracle Database File System

Task 3 - Configure Oracle Database File System

The Database File System (DBFS) is the only recommended solution when configuring Oracle GoldenGate with Oracle Data Guard.

The DBFS user, tablespace, and file system in the database was previously created in the primary database, as detailed in Cloud: Oracle GoldenGate Microservices Architecture on Oracle Exadata Database Service Configuration Best Practices.

Perform the following steps to complete this task:

  • Step 3.1 - Configuring DBFS on Oracle Exadata Database Service
  • Step 3.2 - (PDB Only) Create an Entry in TNSNAMES
  • Step 3.3 - Copy and Edit the mount-dbfs Scripts from the Primary System
  • Step 3.4 - Register the DBFS Resource with Oracle Clusterware

Step 3.1 - Configuring DBFS on Oracle Exadata Database Service

  1. As the opc OS user on the standby system, add the grid user to the fuse group:
    [opc@exastb-node1 ~]$ sudo -u grid
 $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/bin/olsnodes > ~/dbs_group
[opc@exadb-node1 ~]$ dcli -g ~/dbs_group -l opc sudo usermod -a -G fuse grid
  2. As the opc OS user on the standby system, validate that the file /etc/fuse.conf exists and contains the user_allow_other option: 
    [opc@exastb-node1 ~]$ cat /etc/fuse.conf
# mount_max = 1000
user_allow_other
  3. As the opc OS user on the standby system, skip this step if the option user_allow_other is already in the /etc/fuse.conf file. Otherwise run the following commands to add the option: 
    [opc@exastb-node1 ~]$ dcli -g ~/dbs_group -l opc “echo user_allow_other |
 sudo tee -a /etc/fuse.conf”
  4. As the opc OS user on the standby system, create an empty directory that will be used as the mount point for the DBFS filesystem.

    Note:

    It is important that the mount point is identical as the one in the primary system, because the physical location of the Oracle GoldenGate deployment is included within the deployment configuration files. 
    [opc@exastb-node1 ~]$ dcli -g ~/dbs_group -l opc sudo mkdir -p /mnt/dbfs
  5. As the opc OS user on the standby system, change ownership on the mount point directory so the grid OS user can access it:
    [opc@exastb-node1 ~]$ dcli -g ~/dbs_group -l opc
 sudo chown oracle:oinstall /mnt/dbfs

Step 3.2 - (PDB Only) Create an Entry in TNSNAMES

  1. As the oracle OS user on the standby system, add a connect entry in $TNS_ADMIN/tnsnames.ora file. Use the PDB service name created in Step 2.3:
    [oracle@exadb-node1 ~]$ vi $TNS_ADMIN/tnsnames.ora
dbfs =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = IPC)(KEY=LISTENER))
      (CONNECT_DATA =
        (SERVICE_NAME = <PDB_SERVICE_NAME> )
       )
    )
  2. As the oracle OS user, distribute the $TNS_ADMIN/tnsnames.ora file to the rest of the nodes:
    [oracle@exadb-node1 ~]$ /usr/local/bin/dcli -l oracle -g ~/dbs_group
 -f $TNS_ADMIN/tnsnames.ora -d $TNS_ADMIN/

Step 3.3 - Copy and Edit the mount-dbfs Scripts from the Primary System

  1. As the root OS user on the primary system, create a zip file with the files mount-dbfs.conf and mount-dbfs.sh:
    [opc@exapri-node1 ~]$ sudo su -
[root@exapri-node1 ~]# zip -j /tmp/mount-dbfs.zip
 $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/crs/script/mount-dbfs.sh
 /etc/oracle/mount-dbfs.conf
  adding: mount-dbfs.sh (deflated 67%)
  adding: mount-dbfs.conf (deflated 58%)
  2. As the opc OS user on the standby system, copy the mount-dbfs.zip file from the primary system to the standby system:
    [opc@exastb-node1 ~]$ scp exapri-node1.oracle.com:/tmp/mount-dbfs.zip /tmp
  3. As the opc OS user on the standby system, unzip the mount-dbfs.zip file and edit the configuration file mount-dbfs.conf:
    [opc@exastb-node1 ~]$ unzip /tmp/mount-dbfs.zip -d /tmp
Archive:  /tmp/mount-dbfs.zip
  inflating: /tmp/mount-dbfs.sh     
  inflating: /tmp/mount-dbfs.conf 
[opc@exastb-node1 ~]$ vi /tmp/mount-dbfs.conf

    It is recommended that you place them in the same directory as the primary system. You will need to modify the following parameters in the mount-dbfs.conf file to match the standby database:

    • DBNAME
    • TNS_ADMIN
    • PDB_SERVICE
  4. As the opc OS user on the standby system, copy mount-dbfs.conf to the directory /etc/oracle on database nodes and set proper permissions on it:
    [opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc -d /tmp
 -f /tmp/mount-dbfs.conf
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc sudo
 cp /tmp/mount-dbfs.conf /etc/oracle
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc sudo
 chown oracle:oinstall /etc/oracle/mount-dbfs.conf
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc sudo
 chmod 660 /etc/oracle/mount-dbfs.conf
  5. As the opc OS user on the standby system, copy mount-dbfs.sh to the directory $GI_HOME/crs/script on database nodes and set proper permissions on it:
    [opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc sudo mkdir
 $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/crs/script
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l opc sudo chown
 grid:oinstall $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/crs/script
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l grid
 -d $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/crs/script
 -f /tmp/mount-dbfs.sh
[opc@exastb-node1 ~]$ /usr/local/bin/dcli -g ~/dbs_group -l grid chmod 770
 $(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)/crs/script/mount-dbfs.sh

Step 3.3 - Register the DBFS Resource with Oracle Clusterware

When registering the resource with Oracle Clusterware, be sure to create it as a cluster_resource. The reason for using cluster_resource is so the file system can only by mounted on a single node at one time, preventing mounting of DBFS from concurrent nodes creating the potential of concurrent file writes, causing file corruption problems.

If using Oracle Multitenant, make sure to use the service name for the same PDB that contains the DBFS repository as was created in the primary database.

  1. As the grid OS user on the standby system, find the resource name for the database service created in a previous step for the DBFS service dependency: 
    [opc@exastb-node1 ~]$ sudo su - grid
[grid@exastb-node1 ~]$ crsctl stat res |grep <PDB_NAME>
NAME=ora.<DB_UNIQUE_NAME>.<PDB_SERVICE_NAME>.svc
  2. As the oracle OS user on the standby system, register the Clusterware resource by executing the following script:
    [opc@exadb-node1 ~]$ sudo su - oracle
[oracle@exadb-node1 ~]$ vi add-dbfs-resource.sh
##### start script add-dbfs-resource.sh
#!/bin/bash
ACTION_SCRIPT=$(grep ^crs_home /etc/oracle/olr.loc | cut -d=
 -f2)/crs/script/mount-dbfs.sh
RESNAME=dbfs_mount
DEPNAME=ora.<DB_UNIQUE_NAME>.<PDB_SERVICE_NAME>.svc
ORACLE_HOME=$(grep ^crs_home /etc/oracle/olr.loc | cut -d= -f2)
PATH=$ORACLE_HOME/bin:$PATH
export PATH ORACLE_HOME
crsctl add resource $RESNAME \
  -type cluster_resource \
  -attr "ACTION_SCRIPT=$ACTION_SCRIPT, \
         CHECK_INTERVAL=30,RESTART_ATTEMPTS=10, \
         START_DEPENDENCIES='hard($DEPNAME)pullup($DEPNAME)',\
         STOP_DEPENDENCIES='hard($DEPNAME)',\
         SCRIPT_TIMEOUT=300"
##### end script add-dbfs-resource.sh
[oracle@exadb-node1 ~]$ sh add-dbfs-resource.sh

Note:

After creating the $RESNAME resource, in order to stop the $DBNAME database when the $RESNAME resource is ONLINE, you will have to specify the force flag when using srvctl.

For example: srvctl stop database -d $ORACLE_UNQNAME -f