In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-04-05 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Servers >
Share
Shulou(Shulou.com)06/01 Report--
In this issue, the editor will bring you about the DFS and its new features of Windows Server 2008 R2. The article is rich in content and analyzes and describes it from a professional point of view. I hope you can get something after reading this article.
What is a distributed file system (DFS)?
From a naming point of view, the so-called distributed file system refers to storing files in different locations. In the medium-and large-scale enterprise network architecture, some file servers are deployed within the enterprise, specific public folders are deployed on the file server, and shared for internal employees to access. Each employee can connect to the file server through the internal local area network and access the resources under the shared folder in a manner authorized by the administrator.
There are quite a number of enterprises will have some foreign branches, similarly, the employees of these foreign branches also have the need to access the file server. When these employees go to access the file server at the headquarters, they need to cross the WAN link. In general, the WAN link bandwidth is narrow, there may be access delays, packet loss and other problems, the user experience is very poor, but also squeeze the bandwidth. May affect other business communications between the branch office and the headquarters.
In order to improve the user experience and save the WAN link, we can also deploy a server in the branch office, which acts as a copy of the company headquarters file server, and the shared file resources on it are exactly the same as the headquarters server. when users in the branch office access the shared resources, they only need to access the file server located inside the branch office. This is a good solution.
But there are two prominent problems:
1. How to copy the shared file resources of the headquarters file server to the branch office? Manual or automatic?
Obviously, it would be inappropriate to replicate synchronization manually by an administrator.
2. When users are in headquarters and branch offices, the access path (\ server name) must be different because they are accessing different file servers.
For employees who often commute between headquarters and branch offices, it is not only complex but also easy to be confused to accurately master multiple access paths.
The emergence of distributed file system (DFS) has solved this problem. It combines the shared folders on multiple file servers in headquarters and branch offices into a folder group. The design of DFS is based on client / server mode. Any changes in the contents of any shared folder in this folder group will be copied to all other folders, and the replication schedule can be scheduled.
DFS Architecture Topology
Figure 1 DFS architecture topology
Figure 1 shows a typical architecture topology of a distributed file system (DFS) from Contoso. As can be seen from the figure, there are two branch offices, each deploying a Server as a file server. Named NYC-SRV-01 and LON-SRV-01 respectively, these two servers need to install DFS services to synchronize and update resources through the replication function of DFS.
DFS has three advantages:
1. DFS is copied automatically.
DFS replication provides high availability and fault tolerance at the folder level. We can choose to create servers in New York and London as replication groups and set up a shared folder on each server to establish a corresponding replication relationship to maintain data consistency. A replication group is replicated by multiple masters, and data changes on either side are automatically replicated to other members of the replication group. And according to the link between branches, you can set replication scheduling, set the network bandwidth occupied by replication, the time period of replication, and so on.
2. DFS namespace, unified access path
DFS namespace-provides a unified virtual directory for shared folders that are physically located on servers in different locations.
Branch employees in different physical locations have the same path to access shared resources. For example, "\\ Contoso.com\ Marketing" in figure 1, users in both New York and London use this network path to access the resources they need. This name is created as needed in DFS.
There are two categories of namespaces:
Characteristics based on domain independence
Path\\ Domainname\ Namespace\\ Servername\ Namespace
Location active directory and memory cache server registry and memory cache
Size Windows 2000 server mode, up to 5000 folders
Windows 2008 server mode, up to 50000 folders, up to 50000 folders and targets
Namespace server cluster on multiple Availability servers
DFS replication support
3. The principle of visiting nearby
Although employees in different branch offices access the same network path, the process is divided into two steps:
1) user access:\\ contoso.com\ marketing, the client computer contacts the namespace server and receives recommendations
2) cache the recommendation on the client computer, and then contact the * servers recommended in
From this, we can see that the actual access of the user, by default, will be redirected to the local server in the respective branch office by DFS. If the local server fails, DFS redirects the request to another DFS server closest to the user based on the physical location of the user, and the user does not feel the process of being redirected. Because the shared data on the server is synchronized, users don't even know which server they are accessing.
However, it doesn't matter, what matters is that the user sees what he wants to see.
New features of DFS in Windows Server 2008 R2
The three major advantages of DFS have been introduced earlier, but they have already been implemented on the operating system of Windows Server 2003 R2. In Windows Server 2008 R2, DFS has new features.
1. Read-only DFS
In the DFS system shown in the figure below, because DFS is multi-master replication by default, any new, modified or deleted server content in any branch will affect the servers in headquarters and other branch offices. However, some companies hope that the server in the branch office can only read, not modify or delete, and all document updates, modifications and deletions will be synchronized to the branch office after being completed by the headquarters. so as to ensure the controllability and authority of the data.
Read-only replication of Windows Server 2008 R2 has the following characteristics:
One-way replication
Creating, modifying, or deleting local files are not allowed on read-only nodes
Read-only replication folders can coexist with read-write (RW) replication folders
In a read-only copy, new files or modifications to existing files will be rolled back, but not immediately.
Read-only can be dynamically changed to read and write
A read-only replication folder cannot originate from another read-only replication folder
Read-only member: Windows Server 2008 R2
Other members: Windows Server 2003 R2 or newer
2. Remote differential Compression (RDC) replication
DFS replication inter-group replication uses remote differential Compression (RDC) replication, which only replicates the changes of data. It divides the file into many small chunks, replicates only the changed parts, reduces replication data traffic across the WAN, and accelerates synchronization, as shown in the following figure:
RDC replication requirements:
The AD schema must contain DFS replication objects
The server operating system is Windows Server 2003 R2 or Windows Server 2008, and the DFS replication service is installed
The servers in the replication group must belong to the same forest
The replicated folder must be stored in a NTFS volume
For a server cluster, the replication folder must be on the local storage of that node
Antivirus software must be compatible with DFS replication
These are the DFS and its new features of Windows Server 2008 R2 shared by the editor. If you happen to have similar doubts, please refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 0
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.