89 Questions
What is the purpose of creating a Spanner instance?
To allocate resources for Spanner databases
What are the two important choices included in instance creation?
Instance configuration and compute capacity
What does an instance configuration define?
Geographic placement and replication of databases
What does Spanner oversee as a fully managed database service?
Underlying tasks and resources
What does Spanner offer as a free trial instance for 90 days?
Features and capabilities exploration
What is one thing that Spanner does not offer a way to do?
Manually stop or restart a given instance
How can you configure an instance in Google Cloud?
As regional or multi-regional
What does compute capacity define for an instance?
Amount of server and storage resources
When creating an instance, what must you specify if the compute capacity is smaller than 1000 processing units?
Processing units
What is equivalent to 1000 processing units?
1 node
What determines where your data is stored for an instance in Google Cloud?
Regional or multi-regional configuration
When specifying compute capacity, what measurement unit may be used if the capacity is larger than 1000 processing units?
Number of nodes
What does compute capacity define for an instance in Cloud Spanner?
The amount of server and storage resources available to the databases in an instance
When creating an instance with compute capacity smaller than 1000 processing units, what must be used to specify the compute capacity?
Number of processing units
What happens when an instance has fewer than 1000 processing units in Cloud Spanner?
It has limited compute resources and may experience non-linear scaling and performance for some workloads
What is the equivalent of 1000 processing units in Cloud Spanner?
1 node
When specifying compute capacity on an instance in Cloud Spanner, what multiples can be used for processing units?
$100
What should be used to specify larger quantities of compute capacity in Cloud Spanner once the number of processing units reaches 1000?
Multiples of 1000 processing units or as nodes
What is the basis for determining storage limits in Spanner?
Instance's compute capacity
When might Spanner reject writes to a database?
If the database grows to more than 409.6 GB without adding processing units
What does Spanner bill for in terms of instance storage?
Storage utilization
What is the effect of using instances with fewer than 1000 processing units for larger workloads?
Non-linear scaling and intermittent increases in latencies
In what situation can you not decrease compute capacity in Spanner?
When the instance no longer requires so many splits
What happens when you increase the compute capacity of a Spanner instance?
It increases the CPU and RAM resources for each replica
What does Spanner distribute across zones of one or more regions to provide high performance and availability?
Server resources
How can you adjust the compute capacity of a Spanner instance?
By using Google Cloud console, CLI, or client libraries
What does increasing compute capacity do to the server and storage resources in a Spanner instance?
It increases the resources of each replica
What should you monitor when removing compute capacity from a Spanner instance?
CPU utilization and request latencies
What is one thing that Spanner does not have?
Suspend mode
What determines the peak read and write throughput values that a given amount of compute capacity can provide?
Instance configuration, schema design, and dataset characteristics
What defines the geographic placement and replication of the databases in a Cloud Spanner instance?
Instance configuration
What determines the location where data is stored for a Cloud Spanner instance?
Instance configuration
What type of instance configuration spans more than one Google Cloud region?
Multi-region
What is NOT allowed to be changed for base instance configurations in Cloud Spanner?
Replication topology
What is used to specify larger quantities of compute capacity in Cloud Spanner once the number of processing units reaches 1000?
$2^3$
Where are Google Cloud services available for use with Cloud Spanner?
North America, South America, Europe, Asia, and Australia
What is the purpose of the leader region in a multi-region Spanner instance?
Handle all database writes and reduce latency for connecting clients
What is the main role of read-only replicas in a multi-region Spanner instance?
Serve low-latency reads to clients outside of the read-write regions
What is the function of witness regions in a Spanner instance?
Contain a witness replica used to vote on writes in rare events
What is the impact of moving a Spanner instance to a different instance configuration?
Does not cause downtime and Spanner continues to provide transaction guarantees during the move
How does changing the default leader region in a multi-region Spanner instance affect application latency?
Reduces application latency by placing the leader region closer to connecting clients
What is true about read-only replicas in Spanner instances?
Contain read-only data and serve low-latency reads to clients outside of the read-write regions
When is the witness region important in a Spanner instance?
In the rare event that the read-write regions become unavailable
What is one of the benefits of using read-only replicas in a Spanner instance?
Serve low-latency reads to clients outside of the read-write regions
What is true about moving a Spanner instance between regional and multi-region configurations?
Does not cause downtime and continues to provide transaction guarantees during the move
What effect does configuring the default leader region have on database writes and application latency?
Handle all database writes and reduces latency for connecting clients
What is the purpose of choosing a regional instance configuration for your users and services?
To provide the lowest-latency reads and writes for users and services located within a single region
What is the benefit of adding optional read-only replicas to a custom regional instance configuration?
Supporting low-latency stale reads
What is the primary benefit offered by multi-region instances?
Serving low-latency reads from multiple locations close to or within the regions in the configuration
What is the role of witness replicas in a multi-region configuration?
Assuming leadership when the default leader region returns to a healthy state
What do read-write regions in a multi-region configuration contain?
Two read-write replicas, one default leader region, and one witness region
What is the purpose of designating a default leader region in a multi-region configuration?
To contain the database's leader replicas
What factors enable fast writes in a multi-region configuration?
Regions placed geographically close to form a low-latency quorum
What do additional voting replicas in the second read-write region of a multi-region configuration enable?
Voting to commit writes and serving reads
What role do witness regions play in a multi-region instance configuration?
Placing witness replica in a third region
What is the significance of forming a write quorum in Spanner regional configurations?
Ensuring that every mutation requires a write quorum composed of a majority of voting replicas
Why might a multi-region configuration not be optimal in certain situations?
As it can incur additional network latency when replicas communicate with each other to vote on writes
What must be taken into consideration when creating a custom multi-region instance configuration with optional read-only replicas?
The impact on write quorums when adding optional read-only replicas
What does Cloud Spanner use to replicate data at the byte level?
Underlying file system
In Cloud Spanner, what is the high-level organization unit for data replication?
Database splits
What type of replication scheme does Cloud Spanner use for voting on write requests?
Paxos-based
What is the main purpose of creating replicas of each database split in Cloud Spanner?
Geographic locality
In a multi-region Spanner instance, what role do witness regions play?
Write quorum formation
What is the significance of using a synchronous replication scheme in Cloud Spanner?
Data consistency
What is the main responsibility of leader replicas in Google Cloud Spanner?
Handling writes
What is the role of read-only replicas in a Spanner instance?
Supporting only reads
What is a key benefit of using witness replicas in Spanner's multi-region configurations?
Achieving quorums for writes without large storage and compute resources
What is the impact of adding read-only replicas to a custom instance configuration in a different continent than the leader region?
Maintains low read latency in case of unavailability of a read-only replica
In a Spanner instance, what is the role of non-witness replicas in the background during writes?
Log the write and forward it to other eligible replicas
What does client write requests to a Spanner instance always go to first, regardless of the location?
Leader replica
What is true about adding optional read-only replicas to a custom regional instance configuration in Google Cloud Spanner?
Allows scaling reads and supporting low latency stale reads
What is NOT a role of witness replicas in Google Cloud Spanner's multi-region configurations?
Supporting reads and writes
What is one thing that read-only replicas in Spanner instances do NOT vote for?
Leader election for handling reads and writes
What happens when a majority of voting replicas agree to commit a write in Google Cloud Spanner?
The write is committed
What is one key reason for adding witness replicas to a multi-region configuration in Google Cloud Spanner?
Achieving quorums for writes without large storage and compute resources
What is a characteristic of read-write replicas in Google Cloud Spanner?
Support both reads and writes
What is the purpose of adding optional read-only replicas to a custom Spanner instance configuration?
To provide redundancy and improve read performance
What is the main responsibility of leader replicas in Google Cloud Spanner?
To coordinate write operations and maintain consistency
What is the impact of moving a Spanner instance to a different instance configuration?
It may require adjusting the compute capacity and merging splits
What is the main purpose of creating replicas of each database split in Cloud Spanner?
To provide redundancy and fault tolerance
What do read-write regions in a multi-region configuration contain?
Leader and follower replicas
What does compute capacity define for an instance in Cloud Spanner?
The resources available for processing read and write requests
What should you monitor when removing compute capacity from a Spanner instance?
CPU utilization and request latencies
What must your Google Cloud project have in order to increase the compute capacity of an instance?
Sufficient quota
What is required before deleting an instance that has databases with deletion protection enabled?
Disable the deletion protection on all databases
What does Spanner oversee as a fully managed database service?
Its own underlying tasks and resources
What does not offer a way to manually stop or restart a given instance in Spanner?
Spanner
What is the purpose of leader replicas in Google Cloud Spanner?
Ensuring strong consistency and handling distributed transactions
What do labels help organize within Google Cloud resources?
Resources
Study Notes
Spanner Instance Creation
- Creating a Spanner instance involves choosing instance configuration, which defines the geographic placement and replication of databases.
- Two important choices in instance creation are instance configuration and compute capacity.
Instance Configuration
- Instance configuration defines the geographic placement and replication of databases in a Cloud Spanner instance.
- Regional instance configurations span a single region, while multi-region instance configurations span more than one region.
- Multi-region instance configurations are used for high availability and disaster recovery.
Compute Capacity
- Compute capacity defines the server and storage resources for an instance in Cloud Spanner.
- Compute capacity is specified in processing units, with 1000 processing units being equivalent to 1 node.
- When specifying compute capacity larger than 1000 processing units, units of 1000 (e.g., 2000, 3000) can be used.
Instance Management
- Spanner is a fully managed database service, overseeing database storage, replication, and performance.
- Compute capacity can be adjusted for an instance, but decreasing compute capacity is not allowed when there are outstanding writes or reads.
- When removing compute capacity from a Spanner instance, monitor the instance's performance and adjust as needed.
Data Storage
- Data is stored in the region specified in the instance configuration.
- Storage limits are determined by the instance configuration and compute capacity.
- Spanner bills for instance storage based on usage.
Performance and Availability
- Spanner distributes data across zones of one or more regions to provide high performance and availability.
- Read-only replicas can be added to a custom regional instance configuration to improve read performance.
- Moving a Spanner instance to a different instance configuration can affect application latency.
Replication and Voting
- Cloud Spanner uses a synchronous replication scheme to replicate data at the byte level.
- The high-level organization unit for data replication is the database split.
- Witness replicas play a crucial role in forming a write quorum in multi-region configurations.
- Read-only replicas do not vote on write requests.
Leader and Witness Regions
- The leader region is responsible for handling writes and distributing data to read-only replicas.
- Witness regions play a crucial role in forming a write quorum in multi-region configurations.
- Witness replicas do not store data and are used only for voting purposes.
Instance Configuration and Performance
- The default leader region can be changed, affecting application latency.
- Adding optional read-only replicas to a custom regional instance configuration can improve read performance.
- Moving a Spanner instance to a different instance configuration can affect application latency.
Read-Only Replicas
- Read-only replicas can be added to a custom regional instance configuration to improve read performance.
- Read-only replicas do not vote on write requests and do not store data.
- Adding optional read-only replicas to a custom regional instance configuration can improve read performance.
Multi-Region Configurations
- Multi-region configurations span more than one region for high availability and disaster recovery.
- Witness replicas play a crucial role in forming a write quorum in multi-region configurations.
- The primary benefit of multi-region instances is high availability and disaster recovery.
- The main role of read-only replicas in multi-region configurations is to improve read performance.
Witness Replicas
- Witness replicas do not store data and are used only for voting purposes.
- Witness replicas play a crucial role in forming a write quorum in multi-region configurations.
- Witness replicas enable fast writes in multi-region configurations.
Cloud Spanner Benefits
- Cloud Spanner provides high performance, high availability, and disaster recovery capabilities.
- Cloud Spanner uses synchronous replication to replicate data at the byte level.
- Spanner instances can be configured to provide high availability and disaster recovery.
Learn about the primary characteristics of Cloud Spanner instances, including instance configurations and compute capacity. Understand the importance of these choices in determining the location and amount of resources for Spanner databases.
Make Your Own Quizzes and Flashcards
Convert your notes into interactive study material.
Get started for free
