PolarDB for MySQL provides four different editions: Cluster, Multi-primary Cluster, Single Node, and Archive Database. Each edition provides different specifications of compute nodes.
Specifications of compute nodes in different editions
The following sections describe the specifications of compute nodes in different PolarDB for MySQL editions:
- A Cluster or Multi-primary Cluster Edition cluster supports the General-purpose and Dedicated specifications.
- Dedicated: Computing resources such as CPUs that are allocated to each cluster are exclusive to the cluster. This improves the stability and reliability.
- General-purpose: Idle computing resources such as CPUs are shared among clusters on the same server for cost-effectiveness.
- A Single Node cluster uses nodes of the burstable performance type and shares resources in a computing resource pool. This improves resource usage. The Single Node architecture reduces resource costs because no proxy is required.
- A Archive Database cluster supports the Archive Database (High Compression Ratio) and Archive Database Cluster Edition specifications.
- By default, a cluster of the Archive Database (High Compression Ratio) Edition contains one compute node. The compute node is a dedicated node that reduces
the costs incurred on PolarProxy and the overheads of synchronizing redo logs.
Note Archive Database (High Compression Ratio) clusters are unavailable for purchase. However, existing Archive Database (High Compression Ratio) clusters remain available for use. You can update Archive Database (High Compression Ratio) to Archive Database Cluster Edition Dedicated. For more information, see Upgrade an Archive Database Standalone Edition cluster to an Archive Database Cluster Edition cluster.
- An Archive Database Cluster Edition cluster provides a primary node and multiple read-only nodes. The primary node processes
read and write requests, and an Archive Database cluster contains at least one read-only
node. An Archive Database cluster supports the Dedicated and General-purpose specifications.
- Dedicated: Computing resources such as CPUs that are allocated to each cluster are exclusive to the cluster. This improves the stability and reliability.
- General-purpose: Idle computing resources such as CPUs are shared among clusters on the same server for cost-effectiveness.
- By default, a cluster of the Archive Database (High Compression Ratio) Edition contains one compute node. The compute node is a dedicated node that reduces
the costs incurred on PolarProxy and the overheads of synchronizing redo logs.
Specifications of compute nodes
Edition | Node instance type | CPU and memory specifications | Maximum storage capacity | Maximum connections | Maximum inodes in file system1. | Maximum blktags in file system2 | Internal bandwidth | Maximum IOPS (previous PolarProxy editions) | Maximum IOPS (PolarProxy Enterprise Edition) | I/O bandwidth |
---|---|---|---|---|---|---|---|---|---|---|
Multi-primary Cluster (Dedicated) | polar.mysql.mmx4.large | 4 cores and 16 GB of memory | 10 TB | 6,000 | 2,048,000 | 2,560,000 | 10 Gbps | 32,000 | 48,000 | 4 Gbps |
polar.mysql.mmx4.xlarge | 8 cores and 32 GB of memory | 20 TB | 10,000 | 4,096,000 | 5,120,000 | 10 Gbps | 64,000 | 96,000 | 8 Gbps | |
polar.mysql.mmx8.xlarge | 8 cores and 64 GB of memory | 30 TB | 16,000 | 6,144,000 | 7,680,000 | 10 Gbps | 72,000 | 108,000 | 10 Gbps | |
polar.mysql.mmx8.2xlarge | 16 cores and 128 GB of memory | 100 TB | 32,000 | 20,480,000 | 25,600,000 | 10 Gbps | 128,000 | 192,000 | 16 Gbps | |
polar.mysql.mmx8.4xlarge | 32 cores and 256 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.mmx8.8xlarge | 64 cores and 512 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.mmx8.12xlarge | 88 cores and 710 GB of memory | 100 TB | 100,000 | 20,480,000 | 25,600,000 | 25 Gbps | 256,000 | 384,000 | 32 Gbps | |
Multi-primary Cluster (General-purpose) | polar.mysql.mmg2.large | 4 cores and 8 GB of memory | 10 TB | 3,000 | 2,048,000 | 2,560,000 | 10 Gbps | 15,000 | 22500 | 2 Gbps |
polar.mysql.mmg4.large | 4 cores and 16 GB of memory | 10 TB | 6,000 | 2,048,000 | 2,560,000 | 10 Gbps | 21,000 | 31500 | 2 Gbps | |
polar.mysql.mmg2.xlarge | 8 cores and 16 GB of memory | 10 TB | 6,000 | 2,048,000 | 2,560,000 | 10 Gbps | 24,000 | 36,000 | 2 Gbps | |
polar.mysql.mmg4.xlarge | 8 cores and 32 GB of memory | 20 TB | 10,000 | 4,096,000 | 5,120,000 | 10 Gbps | 36,000 | 54,000 | 4 Gbps | |
polar.mysql.mmg4.2xlarge | 16 cores and 64 GB of memory | 30 TB | 16,000 | 6,144,000 | 7,680,000 | 10 Gbps | 42,000 | 63,000 | 4 Gbps | |
polar.mysql.mmg4.4xlarge | 32 cores and 128 GB of memory | 60 TB | 32,000 | 12,288,000 | 15,360,000 | 10 Gbps | 84,000 | 126,000 | 8 Gbps | |
Cluster (Dedicated) | polar.mysql.x4.medium | 2 cores and 8 GB of memory | 5 TB | 6,000 | 1,024,000 | 1,280,000 | 1 Gbps | 8,000 | 12,000 | 1 Gbps |
polar.mysql.x4.large | 4 cores and 16 GB of memory | 10 TB | 8,000 | 2,048,000 | 2,560,000 | 10 Gbps | 32,000 | 48,000 | 4 Gbps | |
polar.mysql.x4.xlarge | 8 cores and 32 GB of memory | 20 TB | 10,000 | 4,096,000 | 5,120,000 | 10 Gbps | 64,000 | 96,000 | 8 Gbps | |
polar.mysql.x8.xlarge | 8 cores and 64 GB of memory | 30 TB | 16,000 | 6,144,000 | 7,680,000 | 10 Gbps | 72,000 | 108,000 | 10 Gbps | |
polar.mysql.x8.2xlarge | 16 cores and 128 GB of memory | 100 TB | 32,000 | 20,480,000 | 25,600,000 | 10 Gbps | 128,000 | 192,000 | 16 Gbps | |
polar.mysql.x8.4xlarge | 32 cores and 256 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.x8.8xlarge | 64 cores and 512 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.x8.12xlarge | 88 cores and 710 GB of memory | 100 TB | 100,000 | 20,480,000 | 25,600,000 | 25 Gbps | 256,000 | 384,000 | 32 Gbps | |
Cluster (General-purpose) | polar.mysql.g2.medium | 2 cores and 4 GB of memory | 5 TB | 1,800 | 1,024,000 | 1,280,000 | 1 Gbps | 4,000 | 6,000 | 1 Gbps |
polar.mysql.g4.medium | 2 cores and 8 GB of memory | 5 TB | 2,000 | 1,024,000 | 1,280,000 | 1 Gbps | 8,000 | 12,000 | 1 Gbps | |
polar.mysql.g2.large | 4 cores and 8 GB of memory | 10 TB | 3,000 | 2,048,000 | 2,560,000 | 10 Gbps | 15,000 | 22,500 | 2 Gbps | |
polar.mysql.g4.large | 4 cores and 16 GB of memory | 10 TB | 6,000 | 2,048,000 | 2,560,000 | 10 Gbps | 21,000 | 31,500 | 2 Gbps | |
polar.mysql.g2.xlarge | 8 cores and 16 GB of memory | 10 TB | 6,000 | 2,048,000 | 2,560,000 | 10 Gbps | 24,000 | 36,000 | 2 Gbps | |
polar.mysql.g4.xlarge | 8 cores and 32 GB of memory | 20 TB | 10,000 | 4,096,000 | 5,120,000 | 10 Gbps | 36,000 | 54,000 | 4 Gbps | |
polar.mysql.g4.2xlarge | 16 cores and 64 GB of memory | 30 TB | 16,000 | 6,144,000 | 7,680,000 | 10 Gbps | 42,000 | 63,000 | 4 Gbps | |
polar.mysql.g4.4xlarge | 32 cores and 128 GB of memory | 100 TB | 32,000 | 2,048,0000 | 25,600,000 | 10 Gbps | N/A | 84,000 | 8 Gbps | |
Single Node | polar.mysql.s2.large | 4 cores and 8 GB of memory | 5 TB | 6,000 | 1,024,000 | 1,280,000 | 1 Gbps | 8,000 | N/A | 1 Gbps |
Archive Database (High Compression Ratio) | polar.mysql.a4.large | 4 cores and 16 GB of memory | 100 TB | 8,000 | 20,480,000 | 25,600,000 | 10 Gbps | 32,000 | N/A | 4 Gbps |
polar.mysql.a4.xlarge | 8 cores and 32 GB of memory | 100 TB | 10,000 | 20,480,000 | 25,600,000 | 10 Gbps | 64,000 | N/A | 8 Gbps | |
polar.mysql.a8.xlarge | 8 cores and 64 GB of memory | 100 TB | 16,000 | 20,480,000 | 25,600,000 | 10 Gbps | 72,000 | N/A | 10 Gbps | |
polar.mysql.a8.2xlarge | 16 cores and 128 GB of memory | 100 TB | 32,000 | 20,480,000 | 25,600,000 | 10 Gbps | 128,000 | N/A | 16 Gbps | |
polar.mysql.a8.4xlarge | 32 cores and 256 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | N/A | 24 Gbps | |
polar.mysql.a8.8xlarge | 64 cores and 512 GB of memory | 100 TB | 64,000 | 20,480,000 | 25,600,000 | 10 Gbps | 192,000 | N/A | 24 Gbps | |
polar.mysql.a8.12xlarge | 88 cores and 710 GB of memory | 100 TB | 100,000 | 20,480,000 | 25,600,000 | 25 Gbps | 256,000 | N/A | 32 Gbps | |
Archive Database Cluster Edition (Dedicated) | polar.mysql.ax4.large | 4 cores and 16 GB of memory | 200 TB | 8,000 | 40,960,000 | 51,200,000 | 10 Gbps | 32,000 | 48,000 | 4 Gbps |
polar.mysql.ax4.xlarge | 8 cores and 32 GB of memory | 200 TB | 10,000 | 40,960,000 | 51,200,000 | 10 Gbps | 64,000 | 96,000 | 8 Gbps | |
polar.mysql.ax8.xlarge | 8 cores and 64 GB of memory | 200 TB | 16,000 | 40,960,000 | 51,200,000 | 10 Gbps | 72,000 | 108,000 | 10 Gbps | |
polar.mysql.ax8.2xlarge | 16 cores and 128 GB of memory | 200 TB | 32,000 | 40,960,000 | 51,200,000 | 10 Gbps | 128,000 | 192,000 | 16 Gbps | |
polar.mysql.ax8.4xlarge | 32 cores and 256 GB of memory | 200 TB | 64,000 | 40,960,000 | 51,200,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.ax8.8xlarge | 64 cores and 512 GB of memory | 200 TB | 64,000 | 40,960,000 | 51,200,000 | 10 Gbps | 192,000 | 288,000 | 24 Gbps | |
polar.mysql.ax8.12xlarge | 88 cores and 710 GB of memory | 200 TB | 100,000 | 40,960,000 | 51,200,000 | 25 Gbps | 256,000 | 384,000 | 32 Gbps | |
Archive Database Cluster Edition (General-purpose) | polar.mysql.ag2.large | 4 cores and 8 GB of memory | 200 TB | 3,000 | 40,960,000 | 51,200,000 | 10 Gbps | 15,000 | 22,500 | 2 Gbps |
polar.mysql.ag4.large | 4 cores and 16 GB of memory | 200 TB | 6,000 | 40,960,000 | 51,200,000 | 10 Gbps | 21,000 | 31,500 | 2 Gbps | |
polar.mysql.ag2.xlarge | 8 cores and 16 GB of memory | 200 TB | 6,000 | 40,960,000 | 51,200,000 | 10 Gbps | 24,000 | 36,000 | 2 Gbps | |
polar.mysql.ag4.xlarge | 8 cores and 32 GB of memory | 200 TB | 10,000 | 40,960,000 | 51,200,000 | 10 Gbps | 36,000 | 54,000 | 4 Gbps | |
polar.mysql.ag4.2xlarge | 16 cores and 64 GB of memory | 200 TB | 16,000 | 40,960,000 | 51,200,000 | 10 Gbps | 42,000 | 63,000 | 4 Gbps | |
polar.mysql.ag4.4xlarge | 32 cores and 128 GB of memory | 200 TB | 32,000 | 40,960,000 | 51,200,000 | 10 Gbps | 84000 | 126,000 | 8 Gbps |
Maximum inodes in file system1 and maximum blktags in file system2:
- PolarDB uses the self-developed distributed file system of Polar File Sytem (PolarFS) to access underlying storage. Similar to traditional file systems, PolarFS contains inodes and blktags. An inode corresponds to a file. A blktag corresponds to a 4M logical space on Elastic Block Storage (EBS). The numbers of inodes and blktags are linearly related to the storage capacity. For example, a 10-TB disk has a maximum of 2,048,000 inodes and 2,560,000 blktags. If either of the maximum capacity, maximum number of inodes, and maximum number of blktags is reached, you must upgrade the specifications.
- Conversion of capacity units:
- 1 TB=1,000 GB
- 1 GB=1,024 MB
- 1 MB=1,024 KB
- 1 KB=1,024 Byte
Note
- The PolarDB clusters that you purchase on January 7, 2022 or later use Enterprise Edition PolarProxy. Their maximum values of input/output operations per second (IOPS), are listed the Maximum IOPS (PolarProxy Enterprise Edition) column. For clusters that you purchase before January 7, 2022, PolarProxy will be gradually upgraded to Enterprise Edition. After upgrade, their maximum IOPS will also become the values listed in the Maximum IOPS (PolarProxy Enterprise Edition). The maximum IOPS is a theoretical value.
- A PolarDB cluster that has 2 CPU cores and 8 GB of memory is sufficient to provide the basic specifications required in tests, trials, and other light-load scenarios. We recommend that you do not use clusters of these specifications in heavy-load production environments. In production environments, we recommend that you use clusters that have at least 8 CPU cores and 32 GB of memory.
- Clusters in a global database network (GDN) do not support specifications of 2 cores and 4 GB of memory or 2 cores and 8 GB of memory.
- You can specify specifications of primary nodes when you create a Cluster Edition cluster. The same node specifications are automatically applied to read-only nodes.
- The maximum number of connections for a cluster varies based on the node specifications of the cluster. If you add nodes to a cluster, the number of connections to the cluster remains unchanged.
- If you require a larger storage capacity of Archive Database (High Compression Ratio) compute nodes, Submit a ticket to increase the maximum storage capacity to 200 TB.