您现在访问的是微软AZURE全球版技术文档网站,若需要访问由世纪互联运营的MICROSOFT AZURE中国区技术文档网站,请访问 https://docs.azure.cn.

教程:装载群集Tutorial: Mount the cluster

本教程讲解如何将 NFS 客户端装载到 Azure FXT Edge Filer 群集。This tutorial teaches you how to mount NFS clients to the Azure FXT Edge Filer cluster. 客户端将会装载添加后端存储时分配的虚拟命名空间路径。Clients mount the virtual namespace paths that you assigned when you added back-end storage.

本教程介绍:This tutorial teaches:

  • 在面向客户端的 IP 地址范围内对客户端进行负载均衡的策略Strategies for load balancing clients across the range of client-facing IP addresses
  • 如何基于面向客户端的 IP 地址和命名空间接合点构造装载路径How to construct a mount path from a client-facing IP address and namespace junction
  • 要在装载命令中使用哪些参数Which arguments to use in a mount command

完成本教程大约需要 45 分钟。This tutorial takes approximately 45 minutes to complete.

装载群集的步骤Steps to mount the cluster

遵循以下步骤将客户端计算机连接到 Azure FXT Edge Filer 群集。Follow these steps to connect client machines to your Azure FXT Edge Filer cluster.

  1. 确定如何在群集节点之间对客户端流量进行负载均衡。Decide how to load-balance client traffic among your cluster nodes. 有关详细信息,请阅读下面的均衡客户端负载Read Balance client load, below, for details.
  2. 确定要装载的群集 IP 地址和接合点路径。Identify the cluster IP address and junction path to mount.
  3. 确定要装载的面向客户端的路径。Determine the client-facing path for the mount.
  4. 使用适当的参数发出装载命令Issue the mount command, with appropriate arguments.

均衡客户端负载Balance client load

为了帮助均衡群集中所有节点之间的客户端请求,应将客户端装载到所有面向客户端的 IP 地址。To help balance client requests among all the nodes in the cluster, you should mount clients to the full range of client-facing IP addresses. 可通过多种方法将此任务自动化。There are several ways to automate this task.

若要了解群集的轮循机制 DNS 负载均衡,请阅读配置 Azure FXT Edge Filer 群集的 DNSTo learn about round-robin DNS load balancing for the cluster, read Configure DNS for the Azure FXT Edge Filer cluster. 若要使用此方法,必须维护一个 DNS 服务器,本系列文章不会介绍此操作。To use this method you must maintain a DNS server, which is not explained in these articles.

适用于小型安装的更简单方法是在装载客户端时使用脚本分配整个范围内的 IP 地址。A simpler method for small installations is to use a script to assign IP addresses throughout the range at client mount time.

其他负载均衡方法可能适用于大型或复杂系统。Other load-balancing methods might be appropriate for large or complicated systems. 请咨询 Microsoft 代表,或提出支持请求以寻求帮助。Consult your Microsoft representative or open a support request for help. (Azure FXT Edge Filer 目前不支持 Azure 负载均衡器。)(Azure Load Balancer is currently not supported with Azure FXT Edge Filer.)

创建装载命令Create the mount command

mount 命令将 Azure FXT Edge Filer 群集上的虚拟服务器 (vserver) 从客户端映射到本地文件系统上的路径。From your client, the mount command maps the virtual server (vserver) on the Azure FXT Edge Filer cluster to a path on the local filesystem.

格式为 mount <FXT cluster path> <local path> {options}The format is mount <FXT cluster path> <local path> {options}

装载命令有三个元素:There are three elements to the mount command:

  • 群集路径 - 下面所述的 IP 地址和命名空间接合点路径的组合cluster path - a combination of IP address and namespace junction path described below
  • 本地路径 - 客户端上的路径local path - the path on the client
  • 装载命令选项 -(使用建议的装载命令选项中已列出)mount command options - (listed in Use recommended mount command options)

创建群集路径Create the cluster path

群集路径是 vserver IP 地址加上命名空间接合点路径的组合 。The cluster path is a combination of the vserver IP address plus the path to a namespace junction. 命名空间接合点是添加存储系统时定义的虚拟路径。The namespace junction is a virtual path that you defined when you added the storage system.

例如,如果使用 /fxt/files 作为命名空间路径,则客户端会将 IP_address:/fxt/files 装载到其本地载入点。For example, if you used /fxt/files as your namespace path, your clients would mount IP_address:/fxt/files to their local mount point.

在命名空间路径字段中使用 /avere/files 添加“添加新交接点”对话框

IP 地址是为 vserver 定义的面向客户端的 IP 地址之一。The IP address is one of the client-facing IP addresses defined for the vserver. 可在群集控制面板中的两个位置找到面向客户端的 IP 范围:You can find the range of client-facing IPs in two places in the cluster Control Panel:

  • VServers 表(仪表板选项卡) -VServers table (Dashboard tab) -

    控制面板的仪表板选项卡,在图表下方的数据表中选择了 VServer 选项卡,并且带有 IP 地址部分

  • 面向客户端的网络设置页面 -Client Facing Network settings page -

    “设置”>“VServer”>“面向客户端的网络”配置页面,在特定虚拟服务器的表的“地址范围”部分周围有一个圆圈

将 IP 地址和命名空间路径相结合可以构成装载命令的群集路径。Combine the IP address and the namespace path to form the cluster path for the mount command.

示例客户端装载命令:mount 10.0.0.12:/sd-access /mnt/fxt {options}Example client mount command: mount 10.0.0.12:/sd-access /mnt/fxt {options}

创建本地路径Create the local path

装载命令的本地路径由你决定。The local path for the mount command is up to you. 可将所需的任何路径结构设置为虚拟命名空间的一部分。You can set any path structure you want as part of the virtual namespace. 设计可方便运行客户端工作流的命名空间和本地路径。Design a namespace and local path that is convenient for your client workflow.

有关面向客户端的命名空间的详细信息,请参阅《群集配置指南》中的命名空间概述For more information about the client-facing namespace, read the Cluster Configuration Guide's namespace overview.

在装载每个客户端时,除了路径以外,还要包含下面所述的装载命令选项In addition to the paths, include the mount command options described below when mounting each client.

要确保无缝装载客户端,请在装载命令中传递这些设置和参数:To ensure a seamless client mount, pass these settings and arguments in your mount command:

mount -o hard,nointr,proto=tcp,mountproto=tcp,retry=30 ${VSERVER_IP_ADDRESS}:/${NAMESPACE_PATH} ${LOCAL_FILESYSTEM_MOUNT_POINT}

必需设置Required settings 说明Description
hard Azure FXT Edge Filer 群集的软装载与应用程序故障和可能的数据丢失相关联。Soft mounts to the Azure FXT Edge Filer cluster are associated with application failures and possible data loss.
proto=netid 此选项支持适当处理 NFS 网络错误。This option supports appropriate handling of NFS network errors.
mountproto=netid 此选项支持对装载操作的网络错误进行适当处理。This option supports appropriate handling of network errors for mount operations.
retry=n 设置 retry=30 以避免瞬时装载失败。Set retry=30 to avoid transient mount failures. (建议在前景装载中使用不同的值。)(A different value is recommended in foreground mounts.)
首选设置Preferred settings 说明Description
nointr 如果客户端使用支持此选项的早期 OS 内核(2008 年 4 月之前),请使用此选项。If your clients use older OS kernels (before April 2008) that support this option, use it. 默认选项为“intr”。The option "intr" is the default.

后续步骤Next steps

装载客户端后,可以测试工作流并开始使用群集。After you have mounted clients, you can test your workflow and get started with your cluster.

如果需要将数据移到新的云核心文件管理器,请通过并行数据引入来利用缓存结构。If you need to move data to a new cloud core filer, take advantage of the cache structure by using parallel data ingest. 将数据移到 vFXT 群集中介绍了一些策略。Some strategies are described in Moving data to a vFXT cluster. (Avere vFXT for Azure 是基于云的产品,它使用的缓存技术非常类似于 Azure FXT Edge Filer。)(Avere vFXT for Azure is a cloud-based product that uses caching technology very similar to the Azure FXT Edge Filer.)

如果需要排查任何硬件问题,请参阅监视 Azure FXT Edge Filer 硬件状态Read Monitor Azure FXT Edge Filer hardware status if you need to troubleshoot any hardware issues.