What are the system requirements for installing an OceanBase server?
The following table lists the minimum system requirements for installing an OceanBase server.
| Server type | Quantity | Minimum functional configuration | Minimum performance configuration |
|---|---|---|---|
| OceanBase Cloud Platform (OCP) server | 1 | 16 cores, 32 GB of memory, and 1.5 TB of storage | 32 cores, 128 GB of memory, 1.5 TB of SSD storage, and 10 Gbit/s NICs |
| OceanBase Database computing server | 3 | 4 cores and 16 GB of memory
NoteThe log disk size must be four times larger than the memory size, and the data disk must be large enough to store target data. |
32 cores, 256 GB of memory, and 10 Gbit/s NICs
Note
|
To ensure high availability for the OCP management service, three management servers are required. You should also implement load balancing through software or hardware, such as F5 or Alibaba Cloud SLB, or utilize the ob_dns software load component provided by the OceanBase database for a three-node deployment.
The following table lists the Linux operating systems that support OceanBase Database.
| Linux operating system | Version | Server architecture |
|---|---|---|
| Alibaba Cloud Linux | 2 | x86_64 or AArch64 |
| AnolisOS | 8.6 and later | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| KylinOS | V10 | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| Unity Operating System (UOS) | V20 | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| NFSChina | V4.0 and later | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| Inspur KOS | V5.8 | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| CentOS/Red Hat Enterprise Linux | V7.x and V8.x | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| SUSE Enterprise Linux Server | 12SP3 and later | x86_64 (including Hygon) |
| Debian | V8.3 and later | x86_64 (including Hygon) |
| openEuler | 20.03 LTS SP1/SP2 and 22.10 LTS | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
| LinxOS | V6.0.99 and V6.0.100 | x86_64 (including Hygon) or AArch64 (Kunpeng and Phytium) |
Note
Before you use an operating system, configure the network and install a package manager such as YUM or Zypper first.
How do I deploy OceanBase Database in the production environment?
The following table describes deployment solutions.
| Solution | Feature | Infrastructure requirement | Scenario |
|---|---|---|---|
| Three replicas in one data center | RPO=0, low RTO, and automatic failover. Resilient to some hardware failures but not to data center or city-wide disasters. | Single data center | No requirements for data center or city-wide disaster recovery capabilities. |
| Three replicas in three data centers in the same region | RPO=0, low RTO, and automatic failover. Resilient to some hardware failures and data center-level disasters, but not city-wide disasters. | Three data centers in the same region. Low network latency between data centers. | Requires data center-level disaster recovery, no city-wide recovery needed. |
| Five replicas in five data centers across three regions | RPO=0, low RTO, and automatic failover. Resilient to some hardware failures, data center-level disasters, and city-wide disasters. | Five data centers across three regions. Two regions must be geographically close to provide low network latency. | Requires both data center-level and city-wide disaster recovery. |
| Two data centers in the same region + Inter-cluster data replication | RPO>0, high RTO, and manual failover required. Resilient to some hardware failures and data center-level disasters, but not city-wide disasters. | Two data centers in the same region. | Two data centers in the same region with data center-level disaster recovery requirements. |
| Five replicas in three data centers across two regions + Inter-cluster data replication | For data center-level failures: RPO=0, low RTO, and automatic failover. For city-wide failures: RPO>0, high RTO, and manual failover required. Resilient to some hardware failures, data center-level, and city-wide disasters. | Three data centers deployed across two regions. | Two cities and three data centers, requiring both data center-level and city-wide disaster recovery. |
What is LSE?
Large System Extensions (LSE) is a feature introduced in ARMv8.1. It provides a set of atomic operations designed to support synchronization and mutual access in multiprocessor environments. These atomic operations include Load-Exclusive (LDE), Store-Exclusive (STE), and Conditional Compare Exchange (CCXE) instructions. LSE also offers new instructions and memory barriers to maintain data consistency and ordering.
Using LSE instructions enables efficient access to shared memory and provides finer-grained locking mechanisms with lower overhead. Compared with traditional synchronization instructions, LSE can reduce lock contention, enhance concurrency performance, and improve scalability.
From which version does OCP support packages with nolse?
OCP started accommodating OceanBase Database RPM packages marked with nolse from V4.3.0, primarily addressing the detection issue of LSE instruction set functionality on ARM architecture.
Considerations for OBD deployment
Both OceanBase Database RPM packages with and without nolse can be uploaded simultaneously. OBD will adaptively deploy according to the system's support. Alternatively, you can upload only the supported packages for installation as needed.
Considerations for OCP Deployment
While both nolse and non-nolse packages can be uploaded to OCP, adaptive installation is not supported. You must select the appropriate package version that supports LSE when deploying the cluster.