Nebula Graph NebulaGraph for Beginners, 20 Jul 2021

Previous Meeting

⏯

youtube image

►

From YouTube: 新手玩转Nebula - Deployment Options (CN Version & EN Subtitle)

Description

在本视频中，Nebula Graph 布道师 Wey 将带你了解 Nebula 的多种部署方式，你的业务是否需要使用容器化部署？单机还是集群？一图带你了解所有，本视频中的 Nebula Graph 部署方式图片你可以通过 Nebula Graph 公众号发送「部署方式」关键词来获得。如果你在 Nebula Graph 部署过程中遇到任何使用问题，欢迎来论坛：https://discuss.nebula-graph.com.cn 和我们交流哟～

A

Hi, hello, everyone, I'm thinking this is as hera.

B

Hera hi, hi, hello, everyone.

A

Today we then went on to talk about the deployment of about nebula. Nebula is.

A

Currently, a wide variety of deployment we want to help. You find your video through this period of deployment case.

A

Then the next bit thinking we ask for everyone brief look at the different deployment.

B

Nebula graph community offers a wide variety of distribution and deployment.

B

How do we choose the most suitable for our way from these deployments.

B

We can do from start to answer these questions if we want to capacity planning to deploy?

B

whether to consider using kubernetes?, stand-alone or cluster deployment?, we use binary packages installed directly? or compiled from the source code?. We are the early adopters to test rapid deployment?.

B

Or to deploy to a production, environment?.

B

If we consider the words of the container of whether we use kubernetes ?, if the answer is yes,, then we have two options: one is using helm chart way to deploy a is that we have an abstract.

B

Crm operator deployment, we look at our team,, considering it, but if you are using docker deployed,, then it we stand-alone or multi-points machine.

B

Is a stand-alone, then, if we use docker compose docker compose is the first time that we are early adopters of the most convenient ways.

B

We can pull up a nebula with docker compose in our notebook cluster.

B

If we want to multi-machine docker deployed,, then we have docker swarm configuration.

B

Can we use in the community inside that if the situation of non-containerized,, then we look at our environment, supports rpm and deb packages?.

B

We environment is x86 architecture?.

B

If it is,, then we can install if these questions are in the negative.

B

Or we prefer to use the source code is compiled using the direct, rpm and deb way. Official distribution, then.

B

We can use the source code to compile way to installation and deployment.

B

If we just ansible, also use this deployment to maintain binary packages,.

B

We can consider nebula ansible manner provided directly to the official use of its ansible package.

B

Above is to choose the way of all the communities we nebula graph distributed.

B

To the deployment process.

A

Xiexie si introduction what about what you said above that different ways of deploying.

A

Hardware requirements are the same?.

B

This is a good question for the hardware requirements and the deployment of fact,. The relationship is not large.

B

But there is a relationship with the purpose of deploying the use of.

B

The test, if we are not to test environment.

B

And test performance,, then generally the minimum 4 cpu 8g of memory is enough to pull up.

B

A cluster, but if it is a production environment,, then computing resources and the amount of data it?

B

is closely linked. So we can refer to the official documentation plan. Accordingly, configuration.

B

Also to be worth noting that the production environment, we recommend the use of nvme ssd as a storage.

A

Oh good, that you mean additional disks can not do.

B

If it is in fact, a production environment such as volume volume us with the local deployment of cloud soundtrack of kubernetes, if.

B

Ssd is certainly the best that can actually, in order to more cloud native, can use ssd, backend cloud disk.

B

But if other environments, such as with dcoker swarm or with bare metal or virtual machine directly to the deployment binary,, then it is.

B

Natural that we should use the local disk, because it was better performance and relatively more natural in that condition,.

B

If it is,, then do the test environment.

B

Such as your test, server, which is the original hdd hard drive,.

B

Or if you want to pull your laptop inside a development or debug environment,, then it is hard to not have mandatory requirements.

B

You can pull up an environment then do understand normal deployment and operation of.

A

The means that if it is to learn the test,, then use hdd hdd is not so much,, but if it is on the production,, then we ssd would suggest.

A

That good for the next question is about the operating system.

A

That we now support nebula are deployed in support of what operating system it?.

A

Then,, what is the difference between mac and windows?.

B

We nebula graph binary theory, can only run on linux, environment,.

B

But if you are testing or debugging the development of.

B

Many modern tools, by means of words you can say you can take advantage of docker desktop in windows and macos inside in a modern operating system running inside a package that.

B

Can provide you with docker compose. It provides a direct pull up a set of environmental.

B

Or you can use some of the example is not uncommon. Will your desktop system inside a virtual windows or linux environment.

B

There are linux son in that environment inside the system that you can run directly up to our nebula graph for testing and debugging purposes, and the use of the.

A

Well, then, I understand that simply means that we support the linux system.

A

As long as either a mac or windows, as long as it runs on linux, systems can be deployed. Nebula,, right?, ~ good because of the ultimate goal, everyone's use, test or.

A

Production ye may all hope well.

A

Then users will often ask me.

A

If it is under the living environment is our three workloads that have placed this graph meta storage,.

A

What needs attention?.

B

This is a good question and in fact our graphd is stateless,.

B

You can go on-demand deployment that storaged. Then we recommend a minimum of three and then you can go unlimited, horizontal expansion to expand its.

B

Metad compare each particular metad raft. One example is the total number of clusters inside.

B

So it must be the singular requirement.

B

That the average amount of data it can meta kept three.

A

Good, understand that I have one last question is used in the end when the production environment.

A

Is k8s is the cloud deployed?.

A

Or use a local machine to deploy it?, then what is the specific difference between them?.

B

This is a good question. Actually, this problem is entirely dependent on the.

B

Operation and maintenance of their colleagues or sre case.

B

For example, if an existing operation and maintenance of the station is completely kubernetes technology,, then.

B

It is natural, they should choose our.

B

Nebula operator, nebula helm or go to kubernetes way to deploy it.

B

In fact, that in general, no major difference.

B

If the existing environment is not k8s for example,, then.

B

You can go is to deploy them both with docker, swarm or binary way.

B

Is mainly to see you, some trade-offs of existing technologies stand.

A

A good understanding, that.

A

Deployment above using nebula is no much difference in two ways.

A

Mainly to see the operation and maintenance of their classmates preference is more inclined to get used to or what kind of . for.

A

Good,, thank you for sharing.

A

That, with regard to the deployment of the current selection to the discussion here,.

A

I hope you understand the deployment nebula graph.

A

And can choose to deploy their own way.

A

Have any questions and then we can leave a message on the video below under bye, bye, ~.