1Z0-078: Oracle Database 19c: RAC, ASM, and Grid Infrastructure AdministrationDemo

For Oracle ASM, ABORT immediately stops the instance without an orderly disk-group dismount. NORMAL waits for in-progress SQL to complete, dismounts disk groups orderly, and waits for connected users to disconnect. IMMEDIATE waits for in-progress SQL but does not wait for users to disconnect; TRANSACTIONAL has the same behavior as IMMEDIATE because an ASM instance has no transactions. Thus description 3 is NORMAL, description 2 can be IMMEDIATE or TRANSACTIONAL, and description 1 is ABORT.

The System Monitor Service (osysmond) runs on every cluster node. The oclumon manage utility can view and change Cluster Health Monitor configuration, including repository retention settings. Oracle documents one ologgerd service per 32 nodes, while Clusterware diagnostic data primarily resides in the Automatic Diagnostic Repository (ADR); diagcollection.pl is run as root from Grid_home/bin.

Oracle Grid Infrastructure installation can create an Oracle ASM disk group to store the Oracle Cluster Registry (OCR), voting files, and the ASM password file. If no ASM disk groups are specified during the installation, the ASM instances are not started when installation completes on the cluster nodes. Oracle ASM is not mandatory for all database files merely because Clusterware files use ASM, and recovery-area disk groups are created separately after Grid Infrastructure installation as needed.

Oracle RAC CDBs can make a PDB available on every instance or only a subset of instances. PDBs can use local temporary tablespaces, and Oracle RAC requires an active undo tablespace for each instance. Shared undo management is not mandatory, and a PDB does not have to use a shared temporary tablespace.

Oracle ACFS supports Oracle Database binaries and database data files, and it supports archive logs. Oracle ACFS does not support Oracle Cluster Registry (OCR) files, voting files, or files associated with the Oracle Grid Infrastructure home. Oracle ACFS overview

Oracle Database QoS Management disables CRSCTL commands that use the policy and policyset nouns, preventing direct Clusterware administration of the policy set. Under user-defined management, a policy set can define different server-pool attributes in separate policies to support changing demand, and it may also define identical attributes in more than one policy. The policy-set scope is explicitly defined by its server-pool names rather than inherently including every server pool in the cluster. Oracle Clusterware Policy-Based Cluster and Capacity Management

OPatch applies the interim patch to the specified database Oracle home on all cluster nodes. In an Oracle RAC environment, OPatch detects the cluster node list and defaults to all-node patching unless an option such as -local restricts it; -local is specifically the option that prevents propagation to other nodes. The command names one database Oracle home, so it does not patch all homes with the same release.

Oracle ADVM extends Oracle ASM by exposing ASM-backed storage through a standard disk-device driver interface. Each ADVM volume is individually named and is used for one file system. Oracle documentation also states that multiple ADVM volume devices can be configured in an ASM disk group, that their file systems can support non-database files, and that COMPATIBLE.ASM and COMPATIBLE.ADVM need only be set to 11.2 to add a volume.

A newly added cluster configuration policy is not activated automatically. Policy activation reconfigures server pools to match the selected policy; nodes can be reassigned between pools, and resources may be stopped or started when the new placement requires it. Oracle documents policy-set operations through CRSCTL, not SRVCTL.

For a frequently used sequence in Oracle RAC, increasing the sequence cache reduces cache replenishment and synchronization activity. NOORDER removes the requirement to generate values in global request order across instances, avoiding the coordination overhead imposed by ORDER.

Oracle ACFS serializes all snapshot operations clusterwide in the kernel. A snapshot can be created from an existing snapshot in the same ACFS file system when the required ADVM compatibility conditions are met. Snapshots are exposed under the mounted original file system and remain online while that file system is mounted. Snapshot management is performed with acfsutil snap commands; read-write snapshots can be changed, so they do not perpetually retain an immutable point-in-time view.

For clients connecting through a SCAN using an Oracle Net connect descriptor, Oracle Net automatically balances connection requests across the SCAN IP addresses. If an initial SCAN IP address is unavailable, the client attempts another returned SCAN address. The SCAN listener redirects the client to the local listener for the appropriate least-loaded instance; the local listener creates the database connection.

An Oracle Clusterware configuration requires at least two network interfaces on each node: a public network interface for client and application-server access, and a private network interface for internode communication (the cluster interconnect). Oracle supports, but does not require, assigning an interface the combined ASM-and-private role.

Oracle recommends the appvipcfg utility to create or delete an application VIP on the default network (ora.net1.network). A VIP may also be added as another Oracle Clusterware-managed resource, so it can be created with crsctl. Application VIPs can fail over to another eligible cluster node; they are not required merely because an application sets DISPLAY, and they do not use the cluster interconnect network.

ASM_POWER_LIMIT sets the rebalance power and can be as high as 1024 when the disk group ASM compatibility is 11.2.0.2 or later. ASM_DISKGROUPS is optional; Oracle ASM supplies suitable defaults for nonrequired parameters, while INSTANCE_TYPE=ASM is the required parameter. ASM_DISKSTRING is dynamically modifiable, and rebalance power does not limit the number of RDBMS instances that access a disk group.

crsctl stop crs stops Oracle High Availability Services on the local server. crsctl disable crs prevents OHAS from starting automatically when that server boots; it applies only to the local server.

A copied encrypted file is encrypted only when the destination is an encrypted directory. Oracle ACFS supports encrypting files and directories in read-write snapshots, and acfsutil encr on provides the -k option to specify the encryption key length for a directory or file. File-system-level decryption requires root or system-administrator privileges, and encryption initialization is distinct from auditing-role setup.

Oracle Clusterware command evaluation can predict the response to resource-management events without making changes. It applies to Clusterware resource types, including cluster resources, local resources, and administrator-defined resources; it does not evaluate database transactions or treat PDBs as these Clusterware resource entities.

Oracle RAC multitenant databases support both policy-managed and administrator-managed services. For a policy-managed deployment, DBCA can create a server pool when the service configuration calls for one. A singleton service is supported for a multitenant database, so it is not prohibited.

For a local PDB clone, the command must run in the CDB root with the commonly granted CREATE PLUGGABLE DATABASE privilege. A source CDB that is not in ARCHIVELOG mode requires the source PDB to be open READ ONLY; RAC instances must use a consistent open mode. Because no FILE_NAME_CONVERT or CREATE_FILE_DEST clause is supplied, Oracle Managed Files or the PDB_FILE_NAME_CONVERT initialization parameter must determine the cloned PDB file names. Oracle documents that the source PDB cannot be closed for this cloning form.

The -p snap001 option creates a child snapshot from an existing snapshot. Oracle ACFS requires the disk group's COMPATIBLE.ADVM attribute to be version 12.1 or later for creation of a snapshot that specifies a parent snapshot.

Grid Naming Service (GNS) is optional because Oracle Grid Infrastructure also supports manual/static IP address configuration when GNS is not enabled. A normal-redundancy ASM disk group that stores Oracle Clusterware files provides three voting files; Oracle ASM manages that count according to the disk-group redundancy level.

Oracle Clusterware removes a failed node from cluster membership and fails over managed resources to available surviving servers. In an Oracle RAC environment, services can fail over and client sessions can be recovered or reconnected on surviving nodes when the service and client failover features are configured.

For Oracle RAC, ADDM can analyze a selected period defined by two AWR snapshots and can perform Database-mode analysis using data from every database instance. RAC-wide analysis also identifies uneven response-time behavior across instances, enabling cluster-level performance diagnosis.