Skip to main content

Key concepts

Key concepts of MySQL service are as follows.

Instance groups

Instance group is object that manages lifecycle of instances and backups.
MySQL instance groups have various states including operational and terminated states, and users can manage instance groups by checking these states.

Instance group lifecycle and states

When creating MySQL instance group, it starts in Pending state and transitions to Available state through Provisioning and Primary Available states. MySQL instance groups in Available state can be accessed through endpoint and port of selected network. Default ports for MySQL instance groups are 3306(Primary) and 3307(Standby), and users can specify desired ports within range of 1024–65535 during creation.

MySQL instance group lifecycle

Image MySQL instance group lifecycle

StateDescription
PendingPreparing to configure MySQL instance group
- When MySQL instance group is first created, it starts in Pending state
ProvisioningCreating instances of MySQL instance group
Primary AvailableAllocating and configuring virtual hardware resources used by Standby instances of MySQL instance group, while remaining available for normal use
* Availability status of Standby instances is displayed together with operational state (availability state) of instance group
* HA configuration itself is maintained and Primary instance remains available
SwitchingProcessing manual failover request or auto failover
AvailableAll instances in MySQL instance group are in Running state and available for normal use
- MySQL instance groups in Available state can be accessed through endpoint IP and port of selected network
- Users can specify ports within range of 1024–65535, and default ports are 3306(Primary), 3307(Standby)
Scaling-OutAdding Standby instances to MySQL instance group
Scaling-InRemoving Standby instances from MySQL instance group
Scaling-UpFlavor of MySQL instance group has been changed (CPU or memory increased)
Scaling-DownFlavor of MySQL instance group has been changed (CPU or memory decreased)
Storage-Full(Warning)Remaining free space in data or log volume is 30 GB or less
Storage-Full(Error)Remaining free space in data or log volume is 5 GB or less
MaintenanceMySQL maintenance operation is in progress
- Operation types: Engine version upgrade, OS kernel upgrade, OS version upgrade, instance restart (shelve/unshelve, reboot, applying IOPS, etc.)
RebootingInstance group state while instance restart is in progress
ModifyingModifying instance group information (such as changing security groups)
ErrorSome instances in MySQL instance group have encountered errors and are temporarily unavailable
TerminatingFully terminating and returning all instances in MySQL instance group, while deleting backups created from instance group
TerminatedBackups created from MySQL instance group have been deleted, and all instances have been terminated and are unavailable

Availability states of instance group

When instance group is in Primary Available state (allocating and configuring Standby resources), operational state (availability state) of instance group and availability state of Standby instances are displayed together.

Instance group stateAvailability stateDescription
PRIMARY-AVAILABLEREADYAll Standby HA configurations are ready
PARTIAL-READYHA configuration is in progress and at least one Standby has been configured
PARTIAL-UNAVAILABLEAt least one Standby has been configured and at least one Standby is unavailable
CONFIGUREDAll Standby configurations are ready
STANDALONENo configured Standby exists and at least one Standby is unavailable
UNAVAILABLEAll Standby instances are unavailable
AVAILABLEREADYAll Standby configurations are ready
-No Standby exists

Instance

Instance is VM (Virtual Machine) managed by instance group.
MySQL instance is virtualized database including CPU and memory, and is object included in MySQL instance group. Each MySQL instance has states and is assigned either Primary or Standby role depending on availability.

Instance roles
Instance roleDescription
Primary      Instance used for general read and write operations
- Users can access instance through provided endpoint (DNS)
StandbyInstance for high availability in preparation for Primary failures
- Created when instance availability is selected as high availability, or can be added from instance list
- Also serves as read-only replica (Read Replica)

Instance types

MySQL service supports optimized instance types according to usage purpose and expected workloads.
Managed database specifications and settings are determined by instance type of MySQL instance group. Instance type can be selected when creating MySQL instance group, and charges apply according to resources used. MySQL instance types optimized for high-performance computing and memory will be added in future.

MySQL instance typeDefault vCPU (count)Memory (GiB)
m2a.large      2      8      
m2a.xlarge416
m2a.2xlarge832
m2a.4xlarge1664
m2a.8xlarge32128
m2a.12xlarge48192
m2a.16xlarge64256
m2a.24xlarge96384
r2a.large216
r2a.xlarge432
r2a.2xlarge864
r2a.4xlarge16128
r2a.8xlarge32256
r2a.12xlarge48384
r2a.16xlarge64512
r2a.24xlarge96768

Instance lifecycle

Instances start in Pending state and transition to Running state through Provisioning and Starting states.

Image MySQL instance lifecycle

Instance states and billing

StateDescriptionBilling
PendingPreparing to configure instanceNot billed
StartingConfiguring database and applying settings to instanceNot billed
RunningInstance is running and available for useBilled
Volume-ExtendingVolume of instance is being extendedBilled
Engine-UpgradeMySQL engine upgrade is in progressBilled
Kernel-UpgradeKernel upgrade is in progressBilled
OS-UpgradeOS upgrade is in progressBilled
RebootInstance is restartingBilled
ResizingInstance type change (scaling up/down) is in progressBilled
Storage-Full(Warning)Remaining free space in instance data or log storage is 30 GB or lessBilled
Storage-Full(Error)Remaining free space in instance data or log storage is 5 GB or lessOnly storage charges apply
ErrorInstance is temporarily unavailable due to error
State updated when DB or API request fails
Not billed
TerminatingInstance is being fully terminated and returnedNot billed
TerminatedInstance has been terminated and is unavailableNot billed

Instance availability configuration and roles

Number of instances and role of each instance are determined according to instance availability configuration.

Instance availability configurationNumber of created instancesDescription
Single (Single, Primary instance)1         Creates only one Primary instance
High availability (HA, Primary and Standby instances)Up to 6Creates Primary and Standby instances
- Creates 1 Primary and 1–5 Standby instances

Instance availability states

Refers to state of data replication between MySQL instances. Applies only to instance groups configured with high availability.

Image MySQL instance availability states

Availability stateDescription
ConfiguredData replication of MySQL instance is functioning normally
ConfiguringConfiguring data replication of MySQL instance
DelayingData replication of MySQL instance is delayed by more than 60 seconds
WarningSyncStandby has more data than Primary
UnavailableData replication of MySQL instance has stopped

Backup

Backups are managed at MySQL instance group level.

Manage backup lifecycle and states

Backups start in Pending state and are completed when they reach Succeeded state through Processing state. Backups created before March 19, 2024 and backups created after that date have different lifecycles (restore, deletion, etc.).

Backup states (created before March 19, 2024)

Image MySQL backup lifecycle, created before March 19, 2024

StateDescription
PendingInitial backup state
ProcessingBackup is currently in progress
In-Use-CheckingChecking whether recovery using successful backup is in progress when deleting backup
Deletion-PendingBackup used for recovery is pending deletion (deletion proceeds after recovery completes)
DeletedBackup has been completely deleted
SucceededBackup completed successfully
DeletingDeleting backup
FatalInternal logic failure state
ErrorBackup terminated abnormally
- State updated when DB or external API request fails in backend logic
Backup states (created after March 19, 2024)

Image MySQL backup lifecycle, created after March 19, 2024

StateDescription
PendingInitial backup state
ProcessingBackup is currently in progress
- Even if deleted in this state, deletion is executed only after backup completes successfully (Succeeded) or fails (Error)
RestoringBackup is being used for restore and cannot be deleted
DeletedBackup has been completely deleted
SucceededBackup completed successfully
DeletingDeleting backup
- If backup is being used for restore, deletion stops and state changes to Restoring
ErrorBackup failed
- Restore unavailable, deletion available

Failover

MySQL supports high availability (HA) feature. When backup instance availability type is selected as high availability, multiple Standby instances can be created, and failover can be performed to minimize service interruption during failures. Both auto failover and manual failover are supported.

caution

Temporary service interruption may occur during failover.

Failover lifecycle

Instance group states during failover

StateDescription
SWITCHINGFailover is in progress. MySQL service is unavailable.
PRIMARY-AVAILABLEConfiguring Standby instances.

Volumes

Databases are provided using volumes, which are Block Storage devices. Therefore, users do not need to directly manage Block Storage such as creating or configuring Block Storage.

Network and security

All MySQL instances provide VPC environment. VPC provides networking and security features, and users can build logically separated service environments according to required network requirements. For more information, see VPC documentation.

Parameter group

You can create and manage MySQL parameter values in groups.

Parameter application status

Image Parameter group application states

StateDescription
PendingWaiting for parameter application.
ApplyingApplying parameters to instance group.
In-SyncParameters have been successfully applied to instance group.
Partial-SyncParameters have been applied only to some instances in instance group.
Occurs when parameter application fails because instances in instance group are not in Failover, Provisioning, or Available state.
If application completes successfully, state changes to In-Sync; if it fails, state changes to Error-Sync.
Error-SyncFailed to apply parameters to cluster. You can retry parameter application. When retried, application state changes to Applying.
SuspendedWaiting for next operation because parameter application failed in preceding instance group during sequential application.
When state changes from Pending to Suspended, you can retry parameter application. When retried, application state changes to Applying.