RMAN on Multitenant DB – Awareness of the Backup Optimization Behavior

Recovery Manager (RMAN) is one of the most popular Oracle databases components with unique Backup/Recovery features. It is fully integrated with the Multitenant Architecture allowing to implement Manage Many-Databases-as-One strategy.

RMAN permits to customize and save several database parameters used during the backup and recovery operations. Such parameters define for example the backup retention policy, the default device type,  how many archivelog copy should be stored, if the backup-sets should be compressed and/or encrypted and so on…

 

Below an example of RMAN setup with the highlight of the parameter CONFIGURE BACKUP OPTIMIZATION ON discussed on the next sections.

RMAN> show all;

RMAN configuration parameters for database with db_unique_name CEFUPRD are:
CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 8 DAYS;
CONFIGURE BACKUP OPTIMIZATION ON;
CONFIGURE DEFAULT DEVICE TYPE TO DISK;
CONFIGURE CONTROLFILE AUTOBACKUP ON;
CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '/BACKUP/Databases/CEFUPRD/%F';
CONFIGURE DEVICE TYPE DISK PARALLELISM 2 BACKUP TYPE TO BACKUPSET;
CONFIGURE DATAFILE BACKUP COPIES FOR DEVICE TYPE DISK TO 1;
CONFIGURE ARCHIVELOG BACKUP COPIES FOR DEVICE TYPE DISK TO 1;
CONFIGURE CHANNEL DEVICE TYPE DISK FORMAT '/BACKUP/Databases/CEFUPRD/%d_%T_%U';
CONFIGURE MAXSETSIZE TO UNLIMITED;
CONFIGURE ENCRYPTION FOR DATABASE OFF;
CONFIGURE ENCRYPTION ALGORITHM 'AES128';
CONFIGURE COMPRESSION ALGORITHM 'BASIC' AS OF RELEASE 'DEFAULT' OPTIMIZE FOR LOAD TRUE;
CONFIGURE RMAN OUTPUT TO KEEP FOR 10 DAYS;
CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED ON ALL STANDBY BACKED UP 1 TIMES TO DISK;
CONFIGURE SNAPSHOT CONTROLFILE NAME TO '/u01/app/oracle/fast_recovery_area/rcocefuprd/cefuprd/snapcf_cefuprd.f';

RMAN>

 

 

Effects of RMAN Backup Optimization ON/OFF

In a Multitenant environment is more important than ever to understand the effects of the parameter CONFIGURE BACKUP OPTIMIZATION which can be set to ON or OFF.

Behavion when set ON

If RMAN determines that a file is identical and it has been backed up, then it is a candidate to be skipped. RMAN must do further checking to determine whether to skip the file, however, because both the retention policy and the backup duplexing feature are factors in the algorithm that determines whether RMAN has sufficient backups on the specified device type. (Definition from Oracle Backup Recovery User’s Guide).

Behavion when set OFF

The RMAN backup always includes all files no matter if they are identical and already backed up within the backup retention window.

 

 

What happens by migrating from Non-CDB to PDB?

Assuming that we have just migrated a non-CDB database to PDB and our pluggable database has 4 tablespaces all open read/write.  The container uses the same RMAN setup included on the top of this page, with CONFIGURE BACKUP OPTIMIZATION ON.

Dispite having a FULL database backup every night, only 1 backup every 8 days will be complete and consistent, because the RMAN backup optimization algorithm will detect the SEED PDB datafiles unchanged and it will skip those files. Therefore if we restore the CDB using the backup-sets generated by one FULL database backup, with no access to the rest of backup-sets inside the retention window, there are great probabilities that the restore will fail.

 

Extract of the CDB backup log which shows that the PDB$SEED datafiles have been skipped because already backed up 1 time during the last 8 days.

RMAN> BACKUP AS COMPRESSED BACKUPSET INCREMENTAL LEVEL = 0 DATABASE PLUS ARCHIVELOG NOT BACKED UP 1 TIMES; 

Starting backup at May 15 2018 00:35:07
current log archived
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=9 instance=clgbprd1 device type=DISK
allocated channel: ORA_DISK_2
channel ORA_DISK_2: SID=168 instance=clgbprd1 device type=DISK
skipping archived logs of thread 1 from sequence 39516 to 39931; already backed up
skipping archived logs of thread 2 from sequence 34457 to 34749; already backed up
channel ORA_DISK_1: starting compressed archived log backup set
channel ORA_DISK_1: specifying archived log(s) in backup set
input archived log thread=2 sequence=34774 RECID=148645 STAMP=976088413
input archived log thread=2 sequence=34775 RECID=148649 STAMP=976088467
input archived log thread=1 sequence=39944 RECID=148655 STAMP=976088552
input archived log thread=2 sequence=34776 RECID=148651 STAMP=976088509
input archived log thread=2 sequence=34777 RECID=148653 STAMP=976088551
input archived log thread=2 sequence=34778 RECID=148657 STAMP=976088700
input archived log thread=1 sequence=39945 RECID=148662 STAMP=976088937
input archived log thread=2 sequence=34779 RECID=148659 STAMP=976088838

...
Starting backup at May 15 2018 00:50:02
using channel ORA_DISK_1
using channel ORA_DISK_2
skipping datafile 2; already backed up 1 time(s)
skipping datafile 4; already backed up 1 time(s)
channel ORA_DISK_1: starting compressed incremental level 0 datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
...

 

Using only the backup-sets above to restore the CDB means that Oracle has to recreate the two skipped datafiles (number 2 and 4) applying the archived logs generated during the initial CDB provisioning.

To note that the full backup starts including archived log from the following sequence:

  • For the Thread 1 – Sequence  39944
  • For the Thread 2 – Sequence  34774

But when Oracle initiates the Media Recovery, it complains because the archived log Thread 1 – Sequence 1 is unavailable:

RMAN> run {
allocate auxiliary channel dsk1 type disk ;
2> allocate auxiliary channel dsk2 type disk ;
allocate auxiliary channel dsk3 type disk ;
allocate auxiliary channel dsk4 type disk ;
3> allocate auxiliary channel dsk5 type disk ;
4> allocate auxiliary channel dsk6 type disk ;
5> duplicate database to 'CEFUAUX' noopen backup location '/BACKUP/Databases/CEFUPRD/backup_20180515_only' nofilenamecheck;
}6> 7> 8> 9>

allocated channel: dsk1
channel dsk1: SID=322 device type=DISK

allocated channel: dsk2
channel dsk2: SID=471 device type=DISK

allocated channel: dsk3
channel dsk3: SID=9 device type=DISK

allocated channel: dsk4
channel dsk4: SID=166 device type=DISK

allocated channel: dsk5
channel dsk5: SID=323 device type=DISK

allocated channel: dsk6
channel dsk6: SID=478 device type=DISK

Starting Duplicate Db at May 15 2018 09:29:15

....

contents of Memory Script:
{
 set until scn 2372623043;
 recover
 clone database
 delete archivelog
 ;
}
executing Memory Script

executing command: SET until clause

Starting recover at May 15 2018 11:24:39

starting media recovery

unable to find archived log
archived log thread=1 sequence=1
Oracle instance started

 

I hope this example helped to understand that while migrating from non-CDB to Multitenant, many Administration tasks should be carefully reviewed due to major architecture changes.

 


 

Advertisements

Oracle 12c – Unified Audit Trail

 

Oracle 12c introduces “Unified Audit Trail” a faster, easier to access and more secure audit system.

It optionally allows to stage the audit records in a dedicated memory buffer (UNIFIED_AUDIT_SGA_QUEUE_SIZE), where they are temporarily grouped before being written into the audit table via batch transactions.

This new audit configuration substantially reduces the transactional overhead generated by the auditing.

 

Important improvements have also done to simplify the utilization:

– One single audit trail for any audit data, in fact  UNIFIED_AUDIT_TRAIL view replaces SYS.AUD$/DBA_AUDIT_TRAIL, SYS.FGA_LOGS$/DBA_FGA_AUDIT_TRAIL, DVSYS.AUDIT_TRAIL$, V$XML_AUDIT_TRAIL and the OS audit files in adump.

– All audit data stored in Oracle secure files.

– Role segregation between:

  • DBA responsible to maintain free space and backup.
  •  AUDIT_ADMIN responsible to manage the audit policies and define the data retention.
  • AUDIT_VIEWER in charge of the reports.

 

Unified Audit Trail introduces also new security options important to mention:

– It is activated with a kernel relink and it doesn’t require additional steps or parameters.

– The new AUDSYS table has a Read-Only Protection for all users. Even the DBA privilege can’t manipulate the audit records!

 

How to activate Unified Audit Trail

--Stop all Oracle processes: databases, listener and Enterprise Manager agent.

--Relink Oracle with the uniaud_on option.
$ cd $ORACLE_HOME/rdbms/lib
$ make -f ins_rdbms.mk uniaud_on ioracle

--Restart all Oracle processes: databases, listener and Enterprise Manager agent.

--Check is Unified Audit Trail is active
SQL> select * from v$option where PARAMETER='Unified Auditing';

INST_ID PARAMETER                       VALUE              CON_ID
------- ------------------------------ ------------------- ----------
 1      Unified Auditing                TRUE                0

 

Optional, but strongly recommended it is possible to relocate the AUDIT segments  from SYSAUX Tablespace to a dedicated one.

SQL> Create tablespace TBS_AUDIT datafile SIZE 2G AUTOEXTEND ON;

BEGIN
DBMS_AUDIT_MGMT.SET_AUDIT_TRAIL_LOCATION(
 audit_trail_type => dbms_audit_mgmt.audit_trail_unified,
 audit_trail_location_value => 'TBS_AUDIT');
END;
/


SQL> select OWNER, SEGMENT_NAME, PARTITION_NAME, SEGMENT_TYPE, BYTES from dba_segments where TABLESPACE_NAME='TBS_AUDIT';

OWNER           SEGMENT_NAME                    PARTITION_NAME                SEGMENT_TYPE       BYTES
--------------- ------------------------------ ------------------------------ ------------------ ----------
AUDSYS           CLI_SWP$1b2a49f1$1$1           HIGH_PART                      TABLE PARTITION   65536
AUDSYS           CLI_SWP$1b2a49f1$1$1           PART_2                         TABLE PARTITION   65536
AUDSYS           CLI_LOB$1b2a49f1$1$1           HIGH_PART                      INDEX PARTITION   65536
AUDSYS           CLI_TIME$1b2a49f1$1$1          HIGH_PART                      INDEX PARTITION   65536
AUDSYS           CLI_LOB$1b2a49f1$1$1           PART_2                         INDEX PARTITION   65536
AUDSYS           CLI_TIME$1b2a49f1$1$1          PART_2                         INDEX PARTITION   65536
AUDSYS           CLI_SCN$1b2a49f1$1$1           PART_2                         INDEX PARTITION   65536
AUDSYS           SYS_IL0000091784C00014$$       SYS_IL_P241                    INDEX PARTITION   65536
AUDSYS           CLI_SCN$1b2a49f1$1$1           HIGH_PART                      INDEX PARTITION   65536
AUDSYS           SYS_IL0000091784C00014$$       SYS_IL_P246                    INDEX PARTITION   65536
AUDSYS           SYS_LOB0000091784C00014$$      SYS_LOB_P244                   LOB PARTITION     131072
AUDSYS           SYS_LOB0000091784C00014$$      SYS_LOB_P239                   LOB PARTITION     131072

12 rows selected.

 

The introduction of Audit Policies have brought flexibility and granularity on what it is possible to audit, here an example using Oracle sys_context function.

CREATE AUDIT POLICY hr_employees
 PRIVILEGES CREATE TABLE
 ACTIONS UPDATE ON HR.EMPLOYEES
 WHEN 'SYS_CONTEXT(''USERENV'', ''SESSION_USER'') != ''HR_ADMIN'''
 EVALUATE PER STATEMENT;

AUDIT POLICY hr_employees;

 

 

 

Bug on Oracle 12c Multitenant & PDB Clone as Snapshot Copy

While automating the refresh of the test databases on Oracle 12c Multitenant environment with ACFS and PDB snapshot copy, I encountered the following BUG:

The column SNAPSHOT_PARENT_CON_ID of the view V$PDBS shows 0 (zero) in case of PDBs created as Snapshot Copy.

This bug prevents to identify the parent-child relationship between a PDB and its own Snapshots Copies.

The test case below explains the problem:

SQL> CREATE PLUGGABLE DATABASE LARTE3SEFU from LARTE3 SNAPSHOT COPY; 
 
 Pluggable database created. 
 
 SQL> select CON_ID, NAME, OPEN_MODE, SNAPSHOT_PARENT_CON_ID from v$pdbs where NAME in ('LARTE3SEFU','LARTE3'); 
 
 CON_ID      NAME          OPEN_MODE  SNAPSHOT_PARENT_CON_ID 
 ---------- -------------- ---------- ---------------------- 
 5          LARTE3         READ ONLY  0 
 16         LARTE3SEFU     MOUNTED    0  <-- This should be 5
 
 2 rows selected. 

A Service Request to Oracle has been opened, I’ll update this post once I have the official answer.

Update from the Service Request: BUG Fixed on version 12.2

How to Create and Clone PDBs

################################################
## How to create a PDB Database from Seed DB  ##
################################################

CREATE PLUGGABLE DATABASE pdb01
  ADMIN USER pdb_adm IDENTIFIED BY <password> ROLES=(DBA)
  PATH_PREFIX = '/u01/'
  STORAGE (MAXSIZE 20G MAX_SHARED_TEMP_SIZE 2048M)
  FILE_NAME_CONVERT = ('+DATA01','+DATA02')
  DEFAULT TABLESPACE users DATAFILE '+DATA02' SIZE 10G AUTOEXTEND ON MAXSIZE 20G
  TEMPFILE REUSE;

ALTER PLUGGABLE DATABASE pdb01 OPEN;  
 


 
##################################################
## How to clone a PDB Database running on ASM   ##
##################################################

ALTER PLUGGABLE DATABASE pdb01 CLOSE;  
ALTER PLUGGABLE DATABASE pdb01 OPEN READ ONLY;

CREATE PLUGGABLE DATABASE pdb02 FROM pdb01;

ALTER PLUGGABLE DATABASE pdb01 OPEN READ WRITE;
ALTER PLUGGABLE DATABASE pdb02 OPEN READ WRITE;

 
 
 
##################################################
## How to clone a PDB Database using ACFS Snapshot Copy
##################################################
 
ALTER PLUGGABLE DATABASE pdb03 CLOSE;
ALTER PLUGGABLE DATABASE pdb03 OPEN READ ONLY;
 
 
CREATE PLUGGABLE DATABASE pdb04 FROM pdb03
FILE_NAME_CONVERT = ('/u03/oradata/CDB2/pdb03/','/u03/oradata/CDB2/pdb04/')
SNAPSHOT COPY;

ALTER PLUGGABLE DATABASE pdb03 CLOSE;
ALTER PLUGGABLE DATABASE pdb03 OPEN READ WRITE;
ALTER PLUGGABLE DATABASE pdb04 OPEN READ WRITE;

Create Multitenant DB

#########################################
##      How to create a CDB Database        ##
###############################################

–The ENABLE PLUGGABLE DATABASE clause defines that this is a Container Database.

CREATE DATABASE cdb_01
USER SYS IDENTIFIED BY <password>
USER SYSTEM IDENTIFIED BY <password>
LOGFILE GROUP 1 ('/u01/logs/redo01a.log','/u02/logs/redo01b.log') SIZE 500M BLOCKSIZE 512,
GROUP 2 ('/u01/logs/redo02a.log','/u02/logs/redo02b.log') SIZE 500M BLOCKSIZE 512,
GROUP 3 ('/u01/logs/redo03a.log','/u02/logs/redo03b.log') SIZE 500M BLOCKSIZE 512
MAXLOGHISTORY 1
MAXLOGFILES 16
MAXLOGMEMBERS 3
MAXDATAFILES 1024
CHARACTER SET AL32UTF8
NATIONAL CHARACTER SET AL16UTF16
EXTENT MANAGEMENT LOCAL
DATAFILE '/u01/cdb_01/system01.dbf' SIZE 1024M REUSE AUTOEXTEND ON NEXT 10240K MAXSIZE UNLIMITED
SYSAUX DATAFILE '/u01/cdb_01/sysaux01.dbf' SIZE 1024M REUSE AUTOEXTEND ON NEXT 10240K MAXSIZE UNLIMITED
DEFAULT TABLESPACE USERS DATAFILE '/u01/cdb_01/users01.dbf' SIZE 50M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED
DEFAULT TEMPORARY TABLESPACE TEMP TEMPFILE '/u01/cdb_01/temp01.dbf' SIZE 500M REUSE AUTOEXTEND ON NEXT 640K MAXSIZE UNLIMITED
UNDO TABLESPACE undotbs1 DATAFILE '/u01/cdb_01/undotbs01.dbf' SIZE 500M REUSE AUTOEXTEND ON NEXT 5120K MAXSIZE UNLIMITED
ENABLE PLUGGABLE DATABASE
SEED
FILE_NAME_CONVERT = ('/u01/cdb_01/','/u01/pdbs/pdbseed/')
SYSTEM DATAFILES SIZE 125M AUTOEXTEND ON NEXT 10M MAXSIZE UNLIMITED
SYSAUX DATAFILES SIZE 100M
USER_DATA TABLESPACE user_data
DATAFILE '/u01/pdbs/pdbseed/user_data.dbf' SIZE 200M REUSE AUTOEXTEND ON MAXSIZE UNLIMITED;

 

Featured

ASM 12c

A powerful framework for storage management

 

1 INTRODUCTION

Oracle Automatic Storage Management (ASM) is a well-known, largely used multi-platform volume manager and file system, designed for single-instance and clustered environment. Developed for managing Oracle database files with optimal performance and native data protection, simplifying the storage management; nowadays ASM includes several functionalities for general-purpose files too.
This article focuses on the architecture and characteristics of the version 12c, where great changes and enhancements of pre-existing capabilities have been introduced by Oracle.
Dedicated sections explaining how Oracle has leveraged ASM within the Oracle Engineered Systems complete the paper.

 

1.1 ASM 12c Instance Architecture Diagram

Below are highlighted the functionalities and the main background components associated to an ASM instance. It is important to notice how starting from Oracle 12c a database can run within ASM Disk Groups or on top of ASM Cluster file systems (ACFS).

 

ASM_db

 

Overview ASM options available in Oracle 12c.

ACFS

 

1.2       ASM 12c Multi-Nodes Architecture Diagram

In a Multi-node cluster environment, ASM 12c is now available in two configurations:

  • 11gR2 like: with one ASM instance on each Grid Infrastructure node.
  • Flex ASM: a new concept, which leverages the architecture availability and performance of the cluster; removing the 1:1 hard dependency between cluster node and local ASM instance. With Flex ASM only few nodes of the cluster run an ASM instance, (the default cardinality is 3) and the database instances communicate with ASM in two possible way: locally or over the ASM Network. In case of failure of one ASM instance, the databases automatically and transparently reconnect to another surviving instance on the cluster. This major architectural change required the introduction of two new cluster resources, ASM-Listener for supporting remote client connections and ADVM-Proxy, which permits the access to the ACFS layer. In case of wide cluster installation, Flex ASM enhances the performance and the scalability of the Grid Infrastructure, reducing the amount of network traffic generated between ASM instances.

 

Below two graphical representations of the same Oracle cluster; on the first drawing ASM is configured with pre-12c setup, on the second one Flex ASM is in use.

ASM architecture 11gR2 like

01_NO_FlexASM_Drawing

 

 

Flex ASM architecture

01_FlexASM_Drawing

 

 

2  ASM 12c NEW FEATURES

The table below summarizes the list of new functionalities introduced on ASM 12c R1

Feature Definition
Filter Driver Filter Driver (Oracle ASMFD) is a kernel module that resides in the I/O

path of the Oracle ASM disks used to validate write I/O requests to Oracle ASM disks, eliminates accidental overwrites of Oracle ASM disks that would cause corruption. For example, the Oracle ASM Filter Driver filters out all non-Oracle I/Os which could cause accidental overwrites.

General ASM Enhancements –       Oracle ASM now replicates physically addressed metadata, such as the disk header and allocation tables, within each disk, offering a better protection against bad block disk sectors and external corruptions.

–       Increased storage limits: ASM can manage up to 511 disk groups and a maximum disk size of 32 PB.

–       New REPLACE clause on the ALTER DISKGROUP statement.

Disk Scrubbing Disk scrubbing checks logical data corruptions and repairs the corruptions automatically in normal and high redundancy disks groups. This process automatically starts during rebalance operations or the administrator can trigger it.
Disk Resync Enhancements It enables fast recovery from instance failure and faster resyncs performance. Multiple disks can be brought online simultaneously. Checkpoint functionality enables to resume from the point where the process was interrupted.
Even Read For Disk Groups If ASM mirroring is in use, each I/O request submitted to the system can be satisfied by more than one disk. With this feature, each request to read is sent to the least loaded of the possible source disks.
ASM Rebalance Enhancements The rebalance operation has been improved in term of scalability, performance, and reliability; supporting concurrent operations on multiple disk groups in a single instance.  In this version, it has been enhanced also the support for thin provisioning, user-data validation, and error handling.
ASM Password File in a Disk Group ASM Password file is now stored within the ASM disk group.
Access Control Enhancements on Windows It is now possible to use access control to separate roles in Windows environments. With Oracle Database services running as users rather than Local System, the Oracle ASM access control feature is enabled to support role separation on Windows.
Rolling Migration Framework for ASM One-off Patches This feature enhances the rolling migration framework to apply oneoff patches released for ASM in a rolling manner, without affecting the overall availability of the cluster or the database

 

Updated Key Management Framework This feature updates Oracle key management commands to unify the key management application programming interface (API) layer. The updated key management framework makes interacting with keys in the wallet easier and adds new key metadata that describes how the keys are being used.

 

 

2.1 ASM 12c Client Cluster

One more ASM functionality explored but still in phase of development and therefore not really documented by Oracle, is ASM Client Cluster

Designed to host applications requiring cluster functionalities (monitoring, restart and failover capabilities), without the need to provision local shared storage.

The ASM Client Cluster installation is available as configuration option of the Grid Infrastructure binaries, starting from version 12.1.0.2.1 with Oct. 2014 GI PSU.

The use of ASM Client Cluster imposes the following pre-requisites and limitations:

  • The existence of an ASM Server Cluster version 12.1.0.2.1 with Oct. 2014 GI PSU, configured with the GNS server with or without zone delegation.
  • The ASM Server Cluster becomes aware of the ASM Client Cluster by importing an ad hoc XML configuration containing all details.
  • The ASM Client Cluster uses the OCR, Voting Files and Password File of the ASM Server Cluster.
  • ASM Client Cluster communicates with the ASM Server Cluster over the ASM Network.
  • ASM Server Cluster provides remote shared storage to ASM Client Cluster.

 

As already mentioned, at the time of writing this feature is still under development and without official documentation available, the only possible comment is that the ASM Client Cluster looks similar to another option introduced by Oracle 12c and called Flex Cluster. In fact, Flex Cluster has the concept of HUB and LEAF nodes; the first used to run database workload with direct access to the ASM disks and the second used to host applications in HA configuration but without direct access to the ASM disks.

 

 

3  ACFS NEW FEATURES

In Oracle 12c the Automatic Storage Management Cluster File System supports more and more types of files, offering advanced functionalities like snapshot, replication, encryption, ACL and tagging.  It is also important to highlight that this cluster file system comply with the POSIX standards of Linux/UNIX and with the Windows standards.

Access to ACFS from outside the Grid Infrastructure cluster is granted by NFS protocol; the NFS export can be registered as clusterware resource becoming available from any of the cluster nodes (HANFS).

Here is an exhaustive list of files supported by ACFS: executables, trace files, logs, application reports, BFILEs, configuration files, video, audio, text, images, engineering drawings, general-purpose and Oracle database files.

The major change, introduced in this version of ACFS, is definitely the capability and support to host Oracle database files; granting access to a set of functionalities that in the past were restricted to customer files only. Among them, the most important is the snapshot image, which has been fully integrated with the database Multitenant architecture, allowing cloning entire Pluggable databases in few seconds, independently from the size and in space efficient way using copy-on-write technology.

The snapshots are created and immediately available in the “<FS_mount_point>.ASFS/snaps” directory, and can be generated and later converted from read-only to read/write and vice versa. In addition, ACFS supports nested snapshots.

 

Example of ACFS snapshot copy:

-- Create a read/write Snapshot copy
[grid@oel6srv02 bin]$ acfsutil snap create -w cloudfs_snap /cloudfs

-- Display Snapshot Info
[grid@oel6srv02 ~]$ acfsutil snap info cloudfs_snap /cloudfs
snapshot name:               cloudfs_snap
RO snapshot or RW snapshot:  RW
parent name:                 /cloudfs
snapshot creation time:      Wed May 27 16:54:53 2015

-- Display specific file info 
[grid@oel6srv02 ~]$ acfsutil info file /cloudfs/scripts/utl_env/NEW_SESSION.SQL
/cloudfs/scripts/utl_env/NEW_SESSION.SQL
flags:        File
inode:        42
owner:        oracle
group:        oinstall
size:         684
allocated:    4096
hardlinks:    1
device index: 1
major, minor: 251,91137
access time:  Wed May 27 10:34:18 2013
modify time:  Wed May 27 10:34:18 2013
change time:  Wed May 27 10:34:18 2013
extents:
-offset ----length | -dev --------offset
0       4096 |    1     1496457216
extent count: 1

--Convert the snapshot from Read/Write to Read-only
acfsutil snap convert -r cloudfs_snap /cloudfs

 --Drop the snapshot 
[grid@oel6srv02 ~]$ acfsutil snap delete cloudfs_snap /cloudfs

Example of Pluggable database cloned using ACFS snapshot copy List of requirements that must be met to use ACFS SNAPSHOT COPY clause:

      • All pluggable database files of the source PDB must be stored on ACFS.

 

 

      • The source PDB cannot be in a remote CDB.

 

 

      • The source PDB must be in read-only mode.

 

 

      • Dropping the parent PDB with the including datafiles clause, does not automatically remove the snapshot dependencies, manual intervention is required.

 

 

SQL> CREATE PLUGGABLE DATABASE pt02 FROM ppq01
2  FILE_NAME_CONVERT = ('/u02/oradata/CDB4/PPQ01/',
3                       '/u02/oradata/CDB4/PT02/')
4  SNAPSHOT COPY;
Pluggable database created.
Elapsed: 00:00:13.70

The PDB snapshot copy imposes few restrictions among which the source database opened in read-only. This requirement prevents the implementation on most of the production environments where the database must remain available in read/write 24h/7. For this reason, ACFS for database files is particularly recommended on test and development where flexibility, speed and space efficiency of the clones are key factors for achieving high productive environment.

Graphical representation of how efficiently create and maintain a Test & Development database environment:

DB_Snapshot

 

 

4 ASM 12c and ORACLE ENGINEERED SYSTEMS

Oracle has developed few ASM features to leverage the characteristics of the Engineered Systems. Analyzing the architecture of the Exadata Storage, we see how the unique capabilities of ASM make possible to stripe and mirror data across independent set of disks grouped in different Storage Cells.

The sections below describe the implementation of ASM on the Oracle Database Appliance (ODA) and Exadata systems.

 

 

4.1 ASM 12c on Oracle Database Appliance

Oracle Database Appliance is a simple, reliable and affordable system engineered for running database workloads. One of the key characteristics present since the first version is the pay-as-you-grow model; it permits to activate a crescendo number of CPU-cores when needed, optimizing the licensing cost. With the new version of the ODA software bundle, Oracle has introduced the configuration Solution-in-a-box; which includes the virtualization layer for hosting Oracle databases and application components on the same appliance, but on separate virtual machines. The next sections highlight how the two configurations are architected and the role played by ASM:

  • ODA Bare metal: available since version one of the appliance, this is still the default configuration proposed by Oracle. Beyond the automated installation process, it is like any other two-node cluster, with all ASM and ACFS features available.

 

ODA_Bare_Metal

 

  • ODA Virtualized: on both ODA servers runs the Oracle VM Server software, also called Dom0. Each Dom0 hosts the ODA Base (or Dom Base), a privileged virtual machine where it is installed the Appliance Manager, Grid Infrastructure and RDBMS binaries. The ODA Base takes advantage of the Xen PCI Pass-through technology to provide direct access to the ODA shared disks presented and managed by ASM. This configuration reduces the VM flexibility; in fact, no VM migration is allowed, but it guarantees almost no I/O penalty in term of performance. After the Dom Base creation, it is possible to add Virtual Machine where running application components. Those optional application virtual machines are also identified with the name of Domain U.

By default, all VMs and templates are stored on a local Oracle VM Server repository, but in order to be able to migrate application virtual machines between the two Oracle VM Servers a shared repository on the ACFS file system should be created.

The implementation of the Solution-in-a-box guarantees the maximum Return on Investment of the ODA, because while licensing only the virtual CPUs allocated to Dom Base, the remaining resources are assigned to the application components as showed on the picture below.

ODA_Virtualized

 

 

4.2 ACFS Becomes the default database storage of ODA

Starting from Version 12.1.0.2, a fresh installation of the Oracle Database Appliance adopts ACFS as primary cluster file system to store database files and general-purpose data. Three file systems are created in the ASM disk groups (DATA, RECO, and REDO) and the new databases are stored in these three ACFS file systems instead of in the ASM disk groups.

In case of ODA upgrade from previous release to 12.1.0.2, all pre-existing databases are not automatically migrated to ACFS; but can coexist with the new databases created on ACFS.

At any time, the databases can be migrated from ASM to ACFS as post upgrade step.

Oracle has decided to promote ACFS as default database storage on ODA environment for the following reasons:

 

  • ACFS provides almost equivalent performance than Oracle ASM disk groups.
  • Additional functionalities on industry standard POSIX file system.
  • Database snapshot copy of PDBs, and NON-CDB version 11.2.0.4 of greater.
  • Advanced functionality for general-purpose files such as replication, tagging, encryption, security, and auditing.

Database created on ACFS follows the same Oracle Managed Files (OMF) standard used by ASM.

 

 

4.3 ASM 12c on Exadata Machine

Oracle Exadata Database machine is now at the fifth hardware generation; the latest software update has embraced the possibility to run virtual environments, but differently from the ODA or other Engineered System like Oracle Virtual Appliance, the VMs are not intended to host application components. ASM plays a key role on the success of the Exadata, because it orchestrates all Storage Cells in a way that appear as a single entity, while in reality, they do not know and they do not talk to each other.

The Exadata, available in a wide range of hardware configurations from 1/8 to multi-racks, offers a great flexibility on the storage setup too. The sections below illustrate what is possible to achieve in term of storage configuration when the Exadata is exploited bare metal and virtualized:

  • Exadata Bare Metal: despite the default storage configuration, which foresees three disk groups striped across all Storage Cells, guaranteeing the best I/O performance; as post-installation step, it is possible to deploy a different configuration. Before changing the storage setup, it is vital to understand and evaluate all associated consequences. In fact, even though in specific cases can be a meaningful decision, any storage configuration different from the default one, has as result a shift from optimal performance to flexibility and workload isolation.

Shown below a graphical representation of the default Exadata storage setup, compared to a custom configuration, where the Storage Cells have been divided in multiple groups, segmenting the I/O workloads and avoiding disruption between environments.

Exa_BareMetal_Disks_Default

Exa_BareMetal_Disks_Segmented.png

  • Exadata Virtualized: the installation of the Exadata with the virtualization option foresees a first step of meticulous capacity planning, defining the resources to allocate to the virtual machines (CPU and memory) and the size of each ASM disk group (DBFS, Data, Reco) of the clusters. This last step is particularly important, because unlike the VM resources, the characteristics of the ASM disk groups cannot be changed.

The new version of the Exadata Deployment Assistant, which generates the configuration file to submit to the Exadata installation process, now in conjunction with the use of Oracle Virtual Machines, permits to enter the information related to multiple Grid Infrastructure clusters.

The hardware-based I/O virtualization (so called Xen SR-IOV Virtualization), implemented on the Oracle VMs running on the Exadata Database servers, guarantees almost native I/O and Networking performance over InfiniBand; with lower CPU consumption when compared to a Xen Software I/O virtualization. Unfortunately, this performance advantage comes at the detriment of other virtualization features like Load Balancing, Live Migration and VM Save/Restore operations.

If the Exadata combined with the virtualization open new horizon in term of database consolidation and licensing optimization, do not leave any option to the storage configuration. In fact, the only possible user definition is the amount of space to allocate to each disk group; with this information, the installation procedure defines the size of the Grid Disks on all available Storage Cells.

Following a graphical representation of the Exadata Storage Cells, partitioned for holding three virtualized clusters. For each cluster, ASM access is automatically restricted to the associated Grid Disks.

Exa_BareMetal_Disk_Virtual

 

 

4.4 ACFS on Linux Exadata Database Machine

Starting from version 12.1.0.2, the Exadata Database Machine running Oracle Linux, supports ACFS for database file and general-purpose, with no functional restriction.

This makes ACFS an attractive storage alternative for holding: external tables, data loads, scripts and general-purpose files.

In addition, Oracle ACFS on Exadata Database Machines supports database files for the following database versions:

  • Oracle Database 10g Rel. 2 (10.2.0.4 and 10.2.0.5)
  • Oracle Database 11g (11.2.0.4 and higher)
  • Oracle Database 12c (12.1.0.1 and higher)

Since Exadata Storage Cell does not support database version 10g, ACFS becomes an important storage option for customers wishing to host older databases on their Exadata system.

However, those new configuration options and flexibility come with one major performance restriction. When ACFS for database files is in use, the Exadata is still not supporting the Smart Scan operations and is not able to push database operations directly to the storage. Hence, for a best performance result, it is recommended to store database files on the Exadata Storage using ASM disk groups.

As per any other system, when implementing ACFS on Exadata Database Machine, snapshots and tagging are supported for database and general-purpose files, while replication, security, encryption, audit and high availability NFS functionalities are only supported with general-purpose files.

 

 

 5 Conclusion

Oracle Automatic Storage Management 12c is a single integrated solution, designed to manage database files and general-purpose data under different hardware and software configurations. The adoption of ASM and ACFS not only eliminates the need for third party volume managers and file systems, but also simplifies the storage management offering the best I/O performance, enforcing Oracle best practices. In addition, ASM 12c with the Flex ASM setup removes previous important architecture limitations:

  • Availability: the hard dependency between the local ASM and database instance, was a single point of failure. In fact, without Flex ASM, the failure of the ASM instance causes the crash of all local database instances.
  • Performance: Flex ASM reduces the network traffic generated among the ASM instances, leveraging the architecture scalability; and it is easier and faster to keep the ASM metadata synchronized across large clusters. Finally yet importantly, only few nodes of the cluster have to support the burden of an ASM instance, leaving additional resources to application processing.

 

Oracle ASM offers a large set of configurations and options; it is now our duty to understand case-by-case, when it is relevant to use one setup or another, with the aim to maximize performance, availability and flexibility of the infrastructure.