AWS Certified SysOps Administrators are highly skilled in deploying, managing and operating various system operations on AWS. A crucial aspect of this role involves managing databases and performing functions such as restore, point-in-time restore, and promoting read replicas for enhanced performance and disaster recovery.
Point-In-Time Restore
Point-in-time restore (PITR) allows administrators to create a new DB instance by restoring data from a specific point in time. This operation can be handy for cases like production failures, user errors, or application bugs.
To perform a point-in-time restore procedure, follow these steps:
- Navigate to the RDS dashboard. Then, in the navigation pane, choose “Snapshots”.
- Click on the “Create Snapshot” button.
- In the “Create DB Snapshot” page, provide the “DB Instance Identifier’ and enter a name for the snapshot in the “Snapshot Name” box.
- Choose “Create”.
To restore the database to a specific time, follow these steps:
- In the RDS dashboard, in the left navigation pane, choose “Instances”.
- Select the DB instance, then choose “Action”.
- Choose “Restore to Point in Time”.
Promote Read Replica
Promoting a Read Replica enables it to function as a standalone DB instance, and it is no longer a Read Replica after the promotion. This technique is particularly useful when you want to offload your primary database traffic or even conduct a disaster recovery operation.
To promote a Read Replica, perform the following steps:
- Navigate to the RDS dashboard.
- In the navigation pane, opt for “Databases”.
- Choose the read replica that you want to promote.
- Opt for “Actions”, then “Promote Read Replica”.
- In the Promote Read Replica panel, choose “Promote Read Replica”.
Comparison: Point-in-Time Restore vs. Promote Read Replica
Here’s a simple comparison table to highlight the essential differences and use cases between point-in-time restore and promoting a Read Replica.
| Feature | Point-In-Time Restore | Promote Read Replica |
|---|---|---|
| Purpose | Restoring DB instance to a specific point in time. | Making Read Replica a stand-alone DB instance. |
| Use Cases | Production failures, user errors, application bugs. | Offload primary DB traffic, disaster recovery. |
| After Action | New DB instance will be created. | Read Replica will no longer be a replica but a standalone DB instance. |
| Source | Can be done from any DB snapshot. | Can only be done on a Read Replica instance. |
To conclude, understanding how to restore databases and promote read replicas on AWS is a vital skill for any AWS Certified SysOps Administrator. It provides better performance, scalability, and allows for effective disaster recovery procedures. The above examples and comparison table give a comprehensive overview of these concepts for anyone preparing for the AWS Certified SysOps Administrator Associate (SOA-C02) exam.
Practice Test
True or False: Point-in-time restores can be performed for any times during your backup retention period.
- True
- False
Answer: True
Explanation: Amazon RDS creates a storage volume snapshot of your DB instance, backing up the entire DB instance and not just individual databases. You can perform a point-in-time recovery on any available backup.
When promoting a read replica, the source DB becomes a standalone DB. Which of the following are true?
- The DB is fully functional after promotion.
- The data of the DB is reset.
- The master DB instance is deleted.
- AWS automatically creates a new read replica to replace the one that was promoted.
Answer: The DB is fully functional after promotion.
Explanation: After a read replica is promoted, it becomes a standalone database that is fully functional and independent from the source DB instance.
True or False: AWS RDS supports promoting read replicas of encrypted DB instances.
- True
- False
Answer: True
Explanation: AWS RDS supports promoting read replicas of encrypted DB instances. The promoted read replica will have the same AWS Key Management Service (KMS) key as the original DB instance.
When would you use a point-in-time restore for a database on RDS?
- To create a new instance that is a copy of the current one.
- To recover data without any data loss.
- To recover a database to the state it was in at any point in the past within the backup retention period.
- To recover specific items from a database.
Answer: To recover a database to the state it was in at any point in the past within the backup retention period.
Explanation: You would use a point-in-time restore to recover a database to the state it was in at any point in the past within your backup retention period.
True or False: You can perform both manual and automatic backups with Amazon RDS.
- True
- False
Answer: True
Explanation: Amazon RDS enables you to set up, operate, and scale relational databases in the cloud, and it does support both automated backups and user-initiated manual backups.
Which of the following are required to promote a read replica?
- A Multi-AZ deployment.
- An available backup of the read replica.
- The read replica should be active and catching up with changes from the source DB instance.
- No outstanding transactions.
Answer: The read replica should be active and catching up with changes from the source DB instance.
Explanation: Before a read replica can be promoted, it should be active and efficiently catching up with the changes that are being made to the source DB instance.
True or False: It is impossible to stop automated backups on Amazon RDS.
- True
- False
Answer: False
Explanation: Automated backups can be stopped on Amazon RDS by setting the retention period to zero. However, this is not always recommended, especially in production environments.
In the event of a database failure, what restoration options does Amazon RDS provide?
- Point-in-time restore.
- Full restore.
- Snapshot restore.
- All of the above.
Answer: All of the above.
Explanation: Amazon RDS provides several ways to restore your data including point-in-time restore, full restore, or restoring from a DB snapshot.
True or False: The restore operation creates a new DB instance with a new endpoint.
- True
- False
Answer: True
Explanation: AWS RDS restore operation does create a new DB instance with a new endpoint. The original database is not modified.
What can be the replication source for a read replica?
- Another read replica.
- An Amazon S3 bucket.
- A source DB instance.
- An Amazon DynamoDB table.
Answer: A source DB instance.
Explanation: A read replica can only replicate from a source DB instance. It cannot replicate from another read replica, Amazon S3 bucket, or Amazon DynamoDB table.
Interview Questions
What is Point-In-Time Restore (PITR) in AWS RDS?
Point-In-Time Restore (PITR) in AWS Relational Database Service (RDS) allows you to create a new DB instance that starts at a specific point in time within your backup retention period.
How can you restore a DB instance to a specified point in time?
You have to use the AWS Management Console, the AWS CLI restore-db-instance-to-point-in-time command, or the Amazon RDS API RestoreDBInstanceToPointInTime operation to restore a DB instance to a specified point in time.
What is the purpose of promoting a read replica in AWS RDS?
Promoting a read replica makes it a standalone database. This lets you have a copy of your data at a specific point in time, which can be useful for testing or for handling production load.
What happens when you promote a read replica?
When you promote a read replica, its status changes to “creating” while the process is in progress and then to “available” when it’s complete.
Is there any impact on the primary database during the promotion of a read replica?
No, promoting a read replica does not impact the performance of your primary database.
What happens to the read replica after it is promoted?
After a read replica is promoted, it functions as a standalone DB instance and is no longer synchronized with the source DB instance.
What is the maximum retention period for automatic backups on AWS RDS?
The maximum retention period for automatic backups on AWS RDS is 35 days.
Can you perform a Point-In-Time Restore (PITR) for a Read Replica?
No, a Point-In-Time Restore (PITR) cannot be performed directly on a Read Replica. However, you can promote a Read Replica to a standalone DB instance and perform PITR.
Can I perform a Point-In-Time Restore (PITR) operation on a Multi-AZ DB instance directly?
Yes, you can perform a Point-In-Time Restore (PITR) operation directly on a Multi-AZ DB instance.
Is there any data loss when you promote a read replica?
No, there is generally no data loss when you promote a read replica as it contains the full set of data up until the point when it was promoted.
Can multiple read replicas be promoted at the same time?
Yes, multiple replicas can be promoted at the same time. This can be useful for horizontal scaling.
Can I see the Point-In-Time Restore (PITR) window in the AWS Management Console?
Yes, you can. Select the DB instance and go to the “Instance Actions” dropdown menu, from there you’ll be able to see the “Restore To Point In Time” option.
What happens if I delete a source DB instance that has active read replicas?
If a source DB instance is deleted, its read replicas are also deleted.
How can you increase your backup retention period in AWS RDS?
You can modify the DB instance and specify a different value for the backup retention period.
Can you choose the precise moment for the Point-In-Time Restore (PITR)?
Yes, but it’s not always exact, the restored DB instance can be any point within the chosen time segment, and not necessarily the exact time that was chosen.
extutorials.com