Huawei Cloud Deployment Manual

1 Introduction¶

1.1 Document Description¶

This document mainly describes the deployment on Huawei Cloud, introducing the complete steps from resource planning and configuration to deploying Guance and running it.

Note:

• This document uses dataflux.cn as the main domain name example. Replace it with the corresponding domain name during actual deployment.

1.2 Key Terms¶

1.3 Deployment Steps Architecture¶

2 Resource List¶

3 Infrastructure Deployment¶

3.1 Deployment Description¶

RDS, Redis, InfluxDB, Elasticsearch should be created according to the configuration requirements, all within the same region under the same VPC network.

3.2 Steps One, Two, Three: CCE Service Creation¶

3.2.1 Step One: Purchase CCE Cluster¶

Go to the official website, select Product in the navigation bar, choose Container, click Cloud Container Engine CCE, and create a CCE cluster.

• It must be in the same region as the RDS, InfluxDB, and Elasticsearch resources created later.
• Choose the cluster scale based on your own situation, then create the worker nodes afterward.

3.2.2 Step Two: Worker Node Configuration¶

Mainly select ECS specifications and quantity. Specifications can be created according to the requirements in the resource list or assessed based on actual conditions, but they cannot be lower than the minimum configuration requirements. The number of worker nodes should be at least 3 or more, and it is recommended that the data disk size should not be less than 100GiB.

3.2.3 Step Three: Plugin Management¶

The everest and coredns plugins are already created by default when purchasing the k8s cluster. Here, it is recommended to add two more plugins for installation. To ensure cluster resources and avoid additional costs, it is suggested to delete the ICAgent plugin.

Install nginx-ingress plugin

1) Configure CPU and memory resource limits.

2) Create Load Balancer ELB.

If needed, you can choose a dedicated elastic load balancer. In this article, we choose the shared type. Note: Choose Elastic Public IP and use an existing one if possible. Alternatively, create a new IP with a bandwidth of 300Mbit/s billed by traffic.

• Differences between dedicated and shared types can be found in the following link:

https://support.huaweicloud.com/productdesc-elb/elb_pro_0004.html

3) Configure nginx-ingress plugin content related to nginx forwarding.

# Add the following content
{
    "allow-backend-server-header": "true",
    "client-header-buffer-size": "32k",
    "enable-underscores-in-headers": "true",
    "forwarded-for-header": "X-Forwarded-For",
    "generate-request-id": "true",
    "ignore-invalid-headers": "true",
    "keep-alive-requests": "100",
    "large-client-header-buffers": "4 32k",
    "proxy-body-size": "20m",
    "proxy-connect-timeout": "20",
    "reuse-port": "true",
    "server-tokens": "false",
    "ssl-protocols": "TLSv1 TLSv1.1 TLSv1.2 SSLv3",
    "ssl-redirect": "false",
    "use-gzip": "true",
    "worker-cpu-affinity": "auto"
}

Install node-local-dns plugin

• Enable DNS Config Injection function.

Delete ICAgent plugin (Optional)

Choose AOM Application Operations Management, uninstall the ICAgent plugin, and turn off the switches for continued log and metrics collection.

3.4 Step Four: Cache Service¶

• You can choose to use the default built-in caching service. If you do not use the default built-in caching service, configure Redis as follows:
• Distributed caching service (Redis version).
• Redis version: 6.0, standard master-slave architecture, supporting dual replicas.
• Configure Redis password.
• Add the ECS internal network IP automatically created by CCE to the Redis whitelist.

3.5 Step Five: GuanceDB for Metrics¶

Installation Documentation

3.6 Step Six: Elasticsearch¶

• Name: Cloud Search Service CSS
• Version: 7.10.2
• Enable security mode to set up administrator account and password.
• Add the ECS internal network IP automatically created by CCE to the Elasticsearch whitelist.

3.7 Step Seven: RDS¶

• Name: Cloud Database RDS
• Version: 8.0, Standard Master-Slave Edition
• Set MySQL root user password.
• Add the ECS internal network IP automatically created by CCE to the MySQL whitelist.

4 kubectl Installation and Configuration¶

4.1 Install kubectl¶

kubectl is a command-line client tool for Kubernetes, which can be used to deploy applications, check and manage cluster resources, etc. Our Launcher is based on this command-line tool for application deployment. For specific installation methods, refer to the official documentation:

Install and Configure kubectl

You can also click on the cluster and find kubectl Click to View.

4.2 Configure kube config¶

Whether to use public or private kubeconfig depends on whether your operations machine has connectivity to the cluster's internal network. Determine the connection method based on the following use cases.

5 Start Installation¶

After completing the operations, refer to the manual Start Installation