Create a Cluster
On the console, find ES Service Elasticsearch under the Big Data category in Products and Services to enter the product page.
Creation
The page will list all the regions and availability zones where the service can be created. After selecting the required availability zone, service version, disk type, node type, disk size, number of nodes, and filling in some initialization information, the service cluster can be quickly deployed.
Cluster Type Selection
**Basic **
Dedicated Master Node
When creating initially, you can choose the cluster type:
* Basic type is the situation where the cluster does not have a separate master node, that is, all nodes are configured as:
`node.master: true`
`node.data: true`
* The dedicated master node type is the situation where the cluster has a separate master node
The configuration of the master node is:
`node.master: true`
`node.data: false`
The configuration of the data node is:
`node.master: false`
`node.data: true`
If your cluster has a large number of nodes and high node configurations, consider using the master node separation type. This ensures that the master node uses a lower configuration to avoid resource waste, and guarantees that the data node uses a higher configuration to improve data processing capabilities. The number of master nodes for the master node separation type is 3, the number of nodes is 5-20, the master node configuration used is 2 cores 8G, and the master node configuration used when the number of nodes is greater than 20 is 4 cores 8G.
It’s worth noting that the basic type does not currently support adding master nodes.
Node Specifications, Node Quantity Selection and Payment Method
Among the selectable node specifications
Classified by disk and specifications:
| Disk type | Identifier |
|---|---|
| RSSD cloud disk | o.es |
| SSD cloud disk | n.es |
Selectable node specifications detailed configuration (refer to the latest node “UES Price” navigation page):
| Model | Name | Configuration |
|---|---|---|
| Memory-intensive | o.es4m.medium | 2 cores 8G |
| Memory-intensive | o.es4m.xlarge | 4 cores 16G |
| Memory-intensive | o.es4m.2xlarge | 8 cores 32G |
| Memory-intensive | o.es4m.4xlarge | 16 cores 64G |
| Regular model | o.es2m.medium | 2 cores 4G |
| Regular model | o.es2m.xlarge | 4 cores 8G |
| Regular model | o.es2m.2xlarge | 8 cores 16G |
| Regular model | o.es2m.4xlarge | 16 cores 32G |
| Regular model | o.es2m.8xlarge | 32 cores 64G |
Compute Specifications
Before choosing the computing specifications, you need to first pay attention to the disk space of the UES node, and then pay attention to the main data stored in the cluster, data increment, etc., to select the node specifications corresponding to the appropriate disk size.
Disk Size
The main factors affecting the storage capacity of Elasticsearch service are:
- Number of replicas: replicas help increase data reliability, but at the same time, it also increases storage costs. The default and recommended number of replicas is 1. For some scenes that can tolerate data loss caused by exceptions, consider setting the number of replicas to 0.
- Data expansion: In addition to the original data, ES needs to store indexes, column storage data, etc. After applying encoding compression and other technologies, it generally expands by 10%.
- Internal overhead: ES occupies about 20% of disk space for segment merging, ES Translog, logs, etc.
- Reserved by the operating system: by default, the operating system reserves 5%, used for key process handling, system recovery, preventing disk fragmentation issues, etc.
- Safety threshold: usually, at least 15% safety threshold is reserved.
Therefore, the actual space occupied by data in Elasticsearch can be estimated using the following formula:
Disk total = Source data × (1 + Number of replicas) × (1 + Data expansion) / (1 - Internal task overhead) / (1 - Operating system reserved) / (1 - Safety threshold)
= Source data × (1 + Number of replicas) × 1.7
Number of Nodes
The choice of node quantity is closely related to index shards, data redundancy, fault transfer, fault recovery, processing efficiency, etc. If you have a deep understanding and proficient use of Elasticsearch, you should be able to choose a reasonable node configuration and quantity; if you are still learning about Elasticsearch, you can choose based on actual business needs and the amount of data to be processed.
To ensure the high availability of the cluster, the basic type cluster defaults to a minimum of 3 nodes, and the master node separation type cluster defaults to 3 master nodes, with a minimum of 2 data nodes.
After successful creation, subsequent clusters can be expanded nodes (does not support master node creation and expansion).
Payment Selection
The billing mode is divided into hourly, monthly, and annual payment. The console supports modifying the billing type.
Set Kibana Account Password
Kibana is an open-source analytics and visualization platform and is a commonly used management tool for Elasticsearch. You can use Kibana to search, view, and interact with index data in the cluster, and use various charts, tables, maps, etc. Kibana can easily display advanced data analysis and visualization.
Kibana makes understanding large amounts of data easy.
The product defaults to enabling the Kibana service and adding access control. You need to set the account password when creating the cluster.
Confirm payment
After the selection and filling in of cluster information is completed, click Create to go to the payment confirmation page, complete the payment and create.
Cluster Deployment
Wait for deployment. The deployment time may vary depending on the scale of the cluster, the deployment time is approximately between 2 and 5 minutes.
During deployment, the cluster status of the cluster in the list will be displayed as “blue” status creating, and after successful creation, the cluster status will dynamically change to “green” status running.
If the creation status remains for a long time, you can click refresh to reacquire the cluster status.