是否有可能有更便宜的5节点集群准备生产？ [英] Is is possible to have a cheaper 5 nodes cluster ready for production?

问题描述

大家好，

首先，谢谢你的帮助。我正在为一家创业公司工作，但还有很多资金，我们需要尽快投入生产。我一直在试着如何降低成本。遵循Microsoft SKU建议
的5节点集群的最低价格将花费我们约600欧元（700美元）。这对我们来说是一笔不小的钱，特别是我们最初会有的流量。

我试图用Bms机器创建一个双节点集群（16gb ssd）但是一旦我升级应用程序，种子节点将运行磁盘空间。

是否有任何配置或技巧（向群集添加额外的存储？）以便我们可以使用A或B机器或不同机器的混合，可以为我们的小负载和顺利升级提供良好的性能？

感谢您的帮助。

种类此致，

/José。

解决方案

何塞何塞。首先，我会偏离使用Service Fabric的B系列。这些是可破坏的机器，因此性能可能会有所不同，对于生产工作负载而言不是一个好的选择。 

接下来，我会指出您的最佳实践文档，以确保您的环境是为生产工作负载正确设置 

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-best-practices-overview

然后，我建议您查看容量规划文档

https：// docs。 microsoft.com/en-us/azure/service-fabric/service-fabric-cluster-capacity

对于生产工作负载，绝对可以使用少于5个节点。话虽这么说，我们不推荐它，因为它不允许你的最高耐久性。但是，如果流量较低且刚刚开始，那可能不是您需要的东西。 

此外，使用较低的耐久性层将决定您可以拥有的节点数

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-cluster-capacity#the-reliability-characteristics-of-the-cluster

如此长的答案很短，你绝对可以为集群使用更少的节点和更小的节点。我们只是为您提供最佳实践文档，以确保可以确定耐用性和性能。 

我将从基本生产要求开始

• 建议将群集主NodeType专用于系统服务，并使用放置约束将应用程序部署到辅助NodeType。
• 推荐的VM SKU是标准D3或标准D3_V2或等效的，至少有14 GB的本地SSD。
• 支持的最小使用VM SKU是标准D1或标准D1_V2或等效的至少14 GB的本地SSD。
• 14-gt lo cal SSD是最低要求。我们的建议至少为50 GB。对于您的工作负载，尤其是在运行Windows容器时，需要更大的磁盘。
• 生产工作负载不支持标准A0等部分核心VM SKU 。
• 出于性能原因，生产工作负载不支持标准A1 SKU。
• 不支持低优先级虚拟机。

正如您将看到的，最小尺寸应该是D1。我不会低于它，因为它会导致性能，耐用性和可靠性问题。 

最后，如果你刚刚开始，你可以考虑运行服务面料的前提。 Service Fabric本身没有实际收费。这是您需要支付的资源。因此，在某些情况下，在本地运行可以节省一些钱。 

希望有帮助，如果您有其他问题请告诉我。 

Hi all,

First of all, thank you for your help. I am working for a startup wihout many funds yet and we need to go to production soon. I keep scratching my head about how to reduce costs. The minimum price for a 5 node cluster following the Microsoft SKU recommendations will cost us about 600 euro montly (700 dollars). Which is a lot of money for us at the moment, specially giving the amount of traffic we will have initially.

I tried to create a two nodes cluster with Bms machines (16gb ssd) but the seed node will run of out disk space once I upgrade the app.

It there any configuration or trick (adding an extra storage to the cluster?) so that we can use A or B machines or mix of different machines that will provide us good performance for the small load we have and smooth upgrades?

I appreciate your help.

Kind Regards,

/ José.

解决方案

Hi Jose. First off, I would stray away from using the B series with Service Fabric. These are burstable machines so the performance can vary and would not be a good choice for a production workload. 

Next, I would point you to our best practices documentation to ensure your environment is setup properly for a production workload 

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-best-practices-overview

Then, I would suggest you review the capacity planning doc

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-cluster-capacity

You can absolutely use less than 5 nodes for a production workload. That being said, we do not recommend it as it does not allow you the top durability possible. But that might not be something you need if you have low traffic and are just starting out. 

Also, using a lower durability tier will determine the number of nodes you can have

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-cluster-capacity#the-reliability-characteristics-of-the-cluster

So long answer short, you can absolutely use less nodes and smaller nodes for your cluster. We simply provide you with best practices documentation to ensure durability and performance can be determined. 

I would start with the basic production requirements

• It's recommended to dedicate your clusters primary NodeType to system services, and use placement constraints to deploy your application to secondary NodeTypes.
• The recommended VM SKU is Standard D3 or Standard D3_V2 or equivalent with a minimum of 14 GB of local SSD.
• The minimum supported use VM SKU is Standard D1 or Standard D1_V2 or equivalent with a minimum of 14 GB of local SSD.
• The 14-GB local SSD is a minimum requirement. Our recommendation is a minimum of 50 GB. For your workloads, especially when running Windows containers, larger disks are required.
• Partial core VM SKUs like Standard A0 are not supported for production workloads.
• Standard A1 SKU is not supported for production workloads for performance reasons.
• Low-priority VMs are not supported.

As you will see, the minimum size you should go with it a D1. I would not going any lower than that as it will cause issues with performance, durability and reliability. 

Lastly, if you are just starting out, you could consider running Service Fabric on Premise. There is no actual charge for Service Fabric itself. It is the resources you need to pay for. So in some cases running locally can save you some money. 

Hope that helps and if you have other questions let me know. 

这篇关于是否有可能有更便宜的5节点集群准备生产？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！