RAC-WhatIS

What Is

1. What is a split brain? "Split brain in Oracle RAC happens when cluster nodes lose communication but keep working independently. This can lead to data conflicts if nodes modify the same data without synchronisation. It's a critical issue caused by network failures or partitions."

2. Split brain scenario example.

Imagine a scenario where you have an Oracle RAC cluster with three nodes: Node A, Node B, and Node C. These nodes are interconnected through a network for communication and data sharing.

Now, let's say there's a temporary network issue that causes Node A to lose connectivity with Node B and Node C. However, Node B and Node C can still communicate with each other.

In this situation, Node A continues to operate independently and may make changes to a specific database block. At the same time, Node B and Node C, unaware of Node A's actions due to the network issue, also modify the same database block independently.

When the network issue resolves and Node A reconnects with the cluster, there's a conflict. The data block that Node A modified independently differs from the version modified by Nodes B and C. This inconsistency in data versions is the essence of a split brain scenario in Oracle RAC.

3. What is Node Eviction?

Node eviction in a cluster is like kicking out a problematic member to keep the group running smoothly. It happens when a node has issues and can't work properly, so it's removed to prevent it from causing problems for the rest of the group.

There would be many common reasons why a node might be evicted from an Oracle RAC cluster:

Network issues: If a node loses connectivity to the cluster interconnect network, it may be evicted from the cluster to prevent it from causing network instability or data corruption.

Hardware failures: If a node experiences a hardware failure, such as a power supply or disk failure, it may be evicted from the cluster to prevent it from causing data corruption or other problems.

Resource constraints: If a node runs out of resources, such as memory or CPU, it may be evicted from the cluster to prevent it from causing performance issues or instability for the other nodes in the cluster.

Software issues: If a node experiences a software issue, such as a hung process or kernel panic, it may be evicted from the cluster to prevent it from causing instability or data corruption.

4. Difference between Crash recovery and Instance recovery? When an instance crashes in a single node database on startup a crash recovery takes place. In a RAC environment, the same recovery for an instance is performed by the surviving nodes called Instance recovery.

5. What is Oracle Local Registry(OLR)?

Contains node-specific information which is required by OHASD.

Every node has its own dedicated OLR file.(it will not be shared between the nodes).

6. What is the use of an OLR file? OLR file present on RAC Local node. While starting the CRS, it wont be able to Access OCR file to find the cluster resource information. Because at this point ASM instance would also be down. So at this point OLR file ( which is present at normal file system) is accessed to start the required resources.

7. What happned, If OLR is missing or corrupted. clusterware can’t be started on that node!As we know, OCRs files are present in the ASM diskgroup, While starting the CRS, it won't be able to Access OCR files to find the cluster resource information in the ASM Diskgroup. Because at this point the ASM instance would also be down. So at this point OLR file (which is present in the normal file system) is accessed to start the required resources.

* The default location of OLR: $GRID_HOME/cdata/*olr

* View the available backups of OLR:$GRID_HOME/bin/ocrconfig -local -showbackup

* Take the backup of OLR manually: (Need to done from ROOT user) -- ocrconfig -local -manualbackup

* Verify integrity of the OLR: cluvfy comp olr

* Check the current OLR location: $GRID_HOME/bin/ocrcheck -local

* View OLR in text format: $GRID_HOME/bin/ocrdump -local -stdout

* Restore OLR from backup: In case OLR is corrupted, Restore it from backup.

8. What is Oracle Notification Service (ONS)?

*ONS provides event-based communication between the database instances and applications, allowing them to receive notifications about various cluster-related events.

Its act in following condition:

* When Node Status Changes: ONS enables applications to receive notifications when the status of a node changes, such as when a node joins or leaves the cluster.

* Service Availability: ONS allows applications to subscribe to service availability events, so they can receive notifications when a service becomes available or unavailable.

* This helps applications to dynamically adjust their connection endpoints based on the availability of services, improving high availability and load balancing.

* Database Failover: In the event of a database instance failure, Oracle RAC automatically performs a failover to another available instance.

ONS plays a vital role in facilitating this failover process by notifying applications about the instance failure and the new primary instance to which they should connect.

Applications can subscribe to failover events and receive notifications to adjust their connection information and continue accessing the database seamlessly.

11. What is a CRS resource?

Oracle clusterware software run by CRS (Cluster Ready Services) by using OCR and Voting disk. Anything that Oracle Clusterware manages is known as a CRS resource. Some examples of CRS resources are database,an instance,a service,a listener,a VIP address,an application process etc.

10. Time Synchronise Time synchronization in Oracle RAC refers to ensuring that all nodes in the cluster have the same system time. It's crucial for maintaining data consistency and avoiding issues like timestamp conflicts or discrepancies in transaction logs.