CN102096678A

CN102096678A - Method for establishing large-capacity network file system memory

Info

Publication number: CN102096678A
Application number: CN2009102423565A
Authority: CN
Inventors: 郭军; 蔡欣欢; 顾兆瑜; 唐娜
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2009-12-09
Filing date: 2009-12-09
Publication date: 2011-06-15

Abstract

The invention discloses a method for establishing a large-capacity network file system (NFS) memory. The method comprises the following steps of: 1, establishing a serial advanced technology attachment (SATA) device memory by using a personal computer (PC); 2, configuring the personal computer as the NFS memory; and 3, accessing the NFS memory by using each operating system. In the method, a PC-SATA technology is used for establishing the NFS memory, so the problem of requirement of temporary storage of large data volume in enterprises is solved, and a PC-SATA-NFS memory can be quickly established under the condition of demand tension of memory devices so as to meet the requirement of temporary storage of large volume. Since the NFS memory has the characteristic of low cost and is flexible and extensible, the storage cost is saved, and the method is practicable to a certain extent and is worthy of popularization in a certain range.

Description

A kind of method of building the witched-capacity network file system memory

Technical field

The present invention relates to the high capacity technical field of memory, relate in particular to a kind of personal computer-serial advanced technology attachment feeder apparatus (PC-SATA) technology of utilizing and build the method for witched-capacity network file system memory, this method is a kind of breakthrough in the present high capacity memory technology, it is innovatively with current serial advanced technology attachment feeder apparatus (Serial Advanced Technology Attachment, SATA) interfacing and high capacity SATA hard disk combine, and by using network file system(NFS) (NetworkFile System, NFS) the network sharing mode has made up mass storage, for enterprises provides interim mass data storage service.

Background technology

At present, network connects storage (Network Attached Storage, NAS), storage area network (Storage Area Network, SAN), Direct Attached Storage (Direct Attached Storage, DAS) and claim current three big main flow storage architectures, also be that most of enterprise uses three kinds of maximum data storage frameworks.

No matter which kind of storage architecture its essence is the read-write operation interface that disk is provided for client (as application server).As shown in Figure 1, in traditional DAS, server directly links to each other with disk array by interfaces such as SCSI, to the use of disk array by server controls; In SAN, replaced the direct line of the SCSI among the DAS, FC with optical fiber switch, promptly replaced shared structure with switching fabric, formed server " multi-to-multi " visit to disk array, but to the use of disk array still by server controls; In NAS, not only replaced shared structure with switching fabric (Ethernet switch), simultaneously file system service is transferred to disk array one deck, then the server resource of need not reallocating is controlled the use of disk array.

1), DAS storage

DAS storage is exactly the built-in disk in the server, all is the built-in disk of SCSI or serial type scsi interface generally, and memory capacity is subjected to the restriction of mainboard SCSI mouth quantity on the server and monolithic disk size.

In the use, DAS stores following advantage:

● use simple, convenient;

● storage speed is faster than common PC disk storage speed;

● after being RIAD, data are more safe and reliable.

The weak point of DAS storage:

● memory capacity is little, the dilatation inconvenience;

● can not carry out the network Sharing Management;

● storage speed is subjected to the restriction of SCSI bus;

● server disk is more expensive than ordinary magnetic disc.

Traditional point-to-point DAS mode can't effectively solve problems such as data sharing and management, and the distance of data transmission also has great restriction, and its limitation has caused the appearance of NAS and SAN.NAS has set up being connected of main frame and memory device by the TCP/IP network, data sharing and centralized management have been realized to a certain extent, and solved the problem of data transmission distance, but its opening and message transmission rate all have bigger restriction, and be easy to cause network congestion, only be applicable to the long Distance Transmission of low volume data.And SAN is the fiber channel network of core by setting up with the optical fiber switch, has solved above problem comprehensively and effectively, and along with the attention and the support of more and more manufacturers, has become the mainstream standard of storage industry.

2), NAS storage

NAS is developed by early stage file server, through years of development, has become based on storage system operating system Ethernet, that have special use and memory management functions, that by the network shared files system kinds of platform is provided data storage service.

NAS is with respect to the advantage of DAS:

● storage networkingization, server and memory device form the switching fabric of " multi-to-multi ";

● data centralization storage, backup are easy to management;

● remote data access;

● support the different platform shared data;

● it is easy to install, upgrade, and production system need not be shut down;

● improved extensibility, can add more memory device according to business demand.

The weak point of NAS:

● data rate is lower;

● back up, resuming operation takies the network bandwidth;

● data are transmitted with file mode, do not support some commercial application, as the database application based on raw device;

● the price comparison costliness.

3), SAN storage

The SAN technology is meant with the optical fiber switch to be core, adopts optical-fibre channel (Fiber Channel) agreement, is the storage networking that transmission medium is set up with optical-fibre channel at a high speed.

With respect to DAS, SAN has following advantage:

● centralized stores, more effectively utilize storage resources, storage speed is the fastest;

● simple, concentrated storage administration, reduced management workload;

● memory device has improved dirigibility and expandability to the multi-to-multi connected mode of server;

● shorten data backup and release time, improved handling capacity;

● improved extensibility, can add more memory device according to business demand;

● need not interrupting service, can add or reconfigure storage resources.

As for data transmission very crucial concerning a lot of application, fiber channel protocol can provide the message transmission rate up to 200MB/s, each PORT also has the bandwidth of 200MB/s on the optical fiber switch, by being connected of main frame and memory device done reasonably planning, can satisfy most demands of using data transmission.

The weak point of SAN storage:

● need build special-purpose fiber optic network, connect optical fiber switch;

● cost an arm and a leg.

At present, SAN storage and NAS storage are the plug-in storages that current each large enterprises mainly adopt, and compare with traditional built-in disk, and SAN stores storage speed faster, and the NAS storage has better flexibility.Simultaneously, use SAN storage and NAS storage that reliable stores service can be provided, some enterprises significant datas have been had safety guarantee preferably.

But the price comparison costliness of above-mentioned two kinds of high-end memory devices is generally all distributed use by planning, all be to use usually in the relatively more crucial place of enterprises, as: that need for a long time or the important data storage of enterprise.If run into the interim high capacity storage demand situation outside the demand planning, just can not obtain corresponding satisfied, enterprise can not store for these ephemeral datas and purchase high-end memory device, can only solve these interim storage demand problems by data migtation or the mode of buying portable disk, so just might influence the ordinary production progress of enterprise.

Thus, we also clearly recognize, along with progressively concentrating and the further raising of production portfolio of application data, how to tackle those and produce interim mass data storage in addition, reduce carrying cost, effectively the management ephemeral data is stored and availability, and optimizes data transmission performance, has become problem demanding prompt solution.

Summary of the invention

(1) technical matters that will solve

In view of this, fundamental purpose of the present invention is to provide a kind of method of building the witched-capacity network file system memory, utilize the PC-SATA technology to build the NFS storer, solved the storage demand problem of the interim big data quantity of enterprises, under the relatively nervous situation of memory device demand, can promptly build the NFS network memory, satisfy interim jumbo memory requirement.Because the dirigibility of this NFS storer, extensibility, characteristics cheaply, both the conserve memory cost had certain practicality again, within the specific limits, was worth of widely use.

(2) technical scheme

For achieving the above object, the invention provides a kind of method of building the witched-capacity network file system memory, this method comprises:

Step 1: utilize personal computer to build serial advanced technology attachment feeder apparatus SATA storage;

Step 2: personal computer is configured to the network file system(NFS) storer.

In the such scheme, described step 1 specifically comprises: select for use the conduct of Suse Linux 10 operating systems to set for this platform, and use the personal computer that has a plurality of SATA II hard-disk interfaces.

In the such scheme, described Suse Linux 10 operating systems carry network file system(NFS) management server function, and this operating system carries software and has the Redundant Array of Independent Disks (RAID) function, can be according to the different demands for security of storage data are selected for use suitable Redundant Array of Independent Disks (RAID) grade.

In the such scheme, described step 2 specifically comprises: land the personal computer that Suse Linux10 operating system is installed with the root identity, editor's share directory configuration file/etc/exports specifies share directory and authority.

In the such scheme, described editor's share directory configuration file/etc/exports specifies share directory and authority, by carrying out with the realization of issuing orders:

1), #vi/etc/exports adds following content in this document, preserve then and withdraw from:

/datapool/share?107.252.78.*(rw，sync，no_root_squash)

The content representation of adding: permission ip address is that the computing machine of 107.252.78.* visits/the datapool/share catalogue with the authority of reading and writing; / datapool/share is also referred to as server output share directory; Parameter meaning in the bracket is described below:

Rw: read/write authority, the parameter of read-only authority are ro;

Sync: data sync write memory and hard disk, or use async, this moment, data can be temporary in the internal memory earlier, and did not write hard disk immediately;

No_root_squash: network file system server share directory attribute of user, if the user is root, share directory just has the authority of root so hereto;

2), carry out as issuing orders its enable port mapping:

#/etc/rc.d/init.d/portmap?start；

3), the startup network file system server of carrying out as issue orders, this moment, network file system(NFS) can activate finger daemon, just began to monitor the request of client then:

#/etc/rc.d/init.d/nfs?start。

In the such scheme, start network file system server in the described step 3) and also can adopt the mode that restarts the Linux server to realize.

In the such scheme, after network file system server starts, also need to check the fire wall setting of Linux server, close firewall services, guarantee not mask network file system(NFS) port that uses and the main frame that allows to communicate by letter, check the setting of option iptables, ipchains in the Linux server, and inspection/etc/hosts.deny ,/etc/hosts.allow file.

In the such scheme, through after the above-mentioned configuration, this method also comprises:

Carry out the loopback test of network file system(NFS) storer on the Linux server, whether the checking share directory can be accessed, specifically is to move on the Linux server as order:

#mount-t nfs 107.252.78.165 (this machine IP) :/datapool/share/mnt

#ls/mnt

Mentioned order is mounted to the network file system(NFS) of Linux server output share directory/the mnt catalogue under, therefore, if the network file system(NFS) operate as normal should be able to be seen in/mnt catalogue/ content in the datapool/share share directory.

In the such scheme, this method also comprises:

Step 3: utilize each operating system access network file system(NFS) storer.

In the such scheme, described each operating system is in (SuSE) Linux OS, AIX operating system, HP-UX operating system and the solaris operating system at least, wherein:

1), utilize (SuSE) Linux OS to connect Linux network file system(NFS) storer down, carry out as issuing orders:

#mount-t?nfs?IP:/share/test；

2), utilize AIX operating system to connect Linux network file system(NFS) storer down, carry out as issuing orders:

#nfso-o?nfs_use_reserved_ports＝1；

#mount?IP:/share/test；

Because the port numbers of using when AIX operating system connects the network file system(NFS) storer with (SuSE) Linux OS is inconsistent, thus before connecting, need to carry out network file system(NFS) memory parameter adjustment operation, otherwise can point out Insufficient privilege, connection failure;

3), utilize HP-UX operating system to connect Linux network file system(NFS) storer down, carry out as issuing orders:

#mount-F?nfs?IP:/share/test；

4), utilize solaris operating system to connect Linux network file system(NFS) storer down, carry out as issuing orders:

#mount-f?nfs-o?vers＝3IP:/share/test。

(3) beneficial effect

Utilize this method of building the witched-capacity network file system memory provided by the invention, the witched-capacity network file system memory of building has been received Expected Results through performance test, compares storage of enterprise-level optical fiber and NAS and stores and have the following advantages:

What 1, dirigibility: PC-SATA-NFS storage was used is PC, so scale also has only the size of common cabinet, and it is smaller to take up an area of the space.Do not need the cloth optical fiber cable,, build very flexible as long as common netting twine is addressable use.

2, the simple and easy storage of extensibility: NFS can also can be changed the bigger SATA hard disk of capacity and reach jumbo purpose according to the SATA mouth quantity dynamic capacity-expanding on the PC.Simultaneously, it also can utilize RAID function among the LINUX that storage is divided into corresponding RAID grade to satisfy different safe class needs.

3, low cost: with compare, the cost of PC-SATA-NFS storage than enterprise-level store cheap tens times to hundred times, power consumption is less than 1/10 of the enterprise-level storage simultaneously, cost performance is higher.

Description of drawings

Fig. 1 is the synoptic diagram of NAS in the prior art, SAN and three kinds of data storage frameworks of DAS;

Fig. 2 is a method flow diagram of building the witched-capacity network file system memory provided by the invention;

Fig. 3 is the structural representation that utilizes the witched-capacity network file system memory that the present invention builds.

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.

At present, a lot of enterprises memory device all service-strong, safety, high speed is deposited significant data, but when running into interim high capacity storage demand with regard to some trouble, because the interim high-end memory device of buying costliness obviously is unpractical, and those interim high capacity storage demands do not need to use so good memory device sometimes, they just need deposit Large Volume Data temporarily, and memory device and security are not had too high requirement.Present enterprises, except above-mentioned high-end memory device, the PC that also has a large amount of low sides, come the jumbo interim storage of fast construction if can make full use of the built-in memory disk of these low side PCs, just can satisfy these interim demands, and improved plant factor, reduced enterprise's unnecessary expenditures.

In order to find cheapness, to build high capacity storage fast and don't that lose performance, we have utilized current 2 important technologies---SATA interfacing and NFS network file system(NFS) technology, proposed to utilize the PC-SATA technology to build the method for NFS storer innovatively, the present invention has not only solved the demand of enterprises temporary needs high capacity storage, and improved the PC-SATA built-in disk and stored and be shared on each different operating system, improve the utilization rate of built-in disk, reduced production cost simultaneously.

1) SATA interfacing and advantage

SATA is the abbreviation of Serial ATA, i.e. serial advanced technology attachment feeder apparatus (serial ATA).This is a kind of novel hard-disk interface type that is different from traditional Parallel ATA fully, owing to adopting serial mode transmission data to gain the name.Compare with Parallel ATA, SATA has bigger advantage.At first, SATA transmits data in the mode of sequential serial, can use higher frequency of operation to improve the bandwidth of data transmission under less bit wide.SATA once only can transmit 1 bit data, can reduce the stitch number of SATA interface like this, and the stube cable number is tailed off, and efficient also can be higher.Secondly, the starting point of SATA is higher, development potentiality is bigger, the data transmission rate of SATA 1.0 definition can reach 150MB/sec, this is also higher than the maximum data transfer rate that the fastest present Parallel ATA (being ATA/133) can reach 133MB/sec, and the data transmission rate of SATA II is then up to 300MB/sec at present.

SATA II grows up on the basis of SATA, to be External Transfer Rate further brought up to 3Gbps (300MB/sec) from the 1.5Gbps (150MB/sec) of SATA to its principal character, comprise in addition the native command formation (Native Command Queuing, NCQ), port Port Multiplier (PortMultiplier), the staggered a series of technical characterictics such as (Staggered Spin-up) that starts.

The gordian technique of SATA II is exactly External Transfer Rate and the NCQ technology of 3Gbps.The NCQ technology can be optimized the instruction execution sequence of hard disk, avoid as conventional hard mechanically according to the diverse location that receives the sequencing moving head reading writing harddisk that instructs, in contrast, it can sort to it after receiving order, magnetic head after the ordering will carry out addressing with high efficiency order, thereby avoid magnetic head to move the loss that brings repeatedly, prolong hard disk life-span.Be not that all SATA hard disks can use the NCQ technology in addition, will support the NCQ, require the SATA controller of board chip set to support NCQ yet except hard disk itself.

The more traditional Parallel ATA wiring of SATAII wiring wants much simple, and folding and unfolding easily, and air-flow in the cabinet and heat radiation are had clear improvement.And the SATAII hard disk is different with the Parallel ATA that is trapped in all the time within the cabinet, and extendibility is very strong, promptly can be external, and the rack of external (JBOD) is not single can to provide better heat radiation and plug function, and more can multiple connection prevent Single Point of Faliure; Because the design of SATAII and optical-fibre channel is exactly the same, so transmission speed can be done assurance with different passages, this is significant in the server and the network storage.

The physical Design of SATAII can be said so with optical-fibre channel (Fibre Channel) as source, so adopt the four-core wiring; The voltage of demand then is reduced to 250mV (the highest 500mV) significantly, few last 200 times of the 5V of more traditional Parallel ATA interface.Therefore, manufacturer can give the senior hard disk function of Serial ATA hard disk affix, as hot plug (Hot Swapping) etc.The more important thing is that on type of attachment, except traditional point-to-point (Point-to-Point) form, SATA also supports " star " to connect, so just can provide facility in the design to the so senior application of RAID; In the use of reality, host bus adaptor (the HBA of SATA, Host Bus Adapter) just look like that switch on the network is the same, can realize with the form of passage and each independent hard disk communication, be that each SATA hard disk is all monopolized a transmission channel, so there is not the problem of master/slave control resemble the Parallel ATA.

At present, maximum monolithic SATA disk can reach the capacity of 1T on the market, and price has only about 600 yuan, compares the built-in scsi disk of server, the SAN storage, and the NAS storage, its price is low-down.Build the simple and easy enterprises that is stored as the storage of interim high capacity is provided if can make full use of the SATA disk, it will save cost for enterprise produces for enterprise brings bigger income.

At present, generally can only the network share directory be provided for other WINDOWS server by the large capacity disc that the built-in SATA disk on the PC is built by the share service that carries in the WINDOWS operating system, certain limitation is arranged in the use, and the performance that the WINDOWS network is shared neither be very good.At this moment, if can utilize a kind of fairly simple and compatible strong, the less network technology of sharing of Network Transmission consumption rate just can build than said method practicality and the outstanding witched-capacity network storage of performance more, we pass through research and experiment, finally selected the sharing mode of NFS, and obtained effect preferably as the high capacity storage.

2) NFS technology brief introduction

Network file system(NFS) (Network File System, NFS) be a kind of a kind of mechanism that subregion on the distance host (catalogue) is mounted to local system through network, by the support to network file system(NFS), the user can come the shared partition (catalogue) to distance host to operate as the local subregion of operation on local system.

NFS is developed out by Sun Microsystems.The function of its maximum is exactly to pass through network, allows different machines, different operating system, can share file separately each other, so, also can simply it be regarded as a File Server.The catalogue that this NFS Server can allow PC that the NFS main frame of network remote is shared is mounted in the middle of the machine of local side, seem at the machine of local side, the catalogue of that distance host just look like be own Partition (subregion or catalogue) generally.

Utilize the NFS technology, we just can with in join the synthetic big storage of disk groups, share and use for different operating system, reach good compatibility, improved the usable range of high capacity storage.

As shown in Figure 2, Fig. 2 is a method flow diagram of building the witched-capacity network file system memory provided by the invention, and this method may further comprise the steps:

In this step, select for use the conduct of Suse Linux 10 operating systems to set for this platform, and use the personal computer (being PC) that has a plurality of SATA II hard-disk interfaces.The data transmission rate of SATA II is up to 300MB/sec at present, and the present invention uses the PC that has four SATAII interfaces, can connect four SATAII hard disks altogether.The most frequently used capacity of SATAII hard disk be 500GB to 1TB, promptly the built-in hard disk of 1 PC can reach the capacity of 2TB to 4TB altogether, fully at last small-sized a storage.

Consider that for the legal operating system software and the convenience of building nfs server the present invention selects for use Suse Linux 10 operating systems as setting for this platform.Suse Linux 10 has carried NFS management server function, and this operating system carries software and have the Redundant Array of Independent Disks function, can be according to the different demands for security of storage data are selected for use suitable R AID grade.

Step 2: personal computer is configured to the network file system(NFS) storer;

Land the above-mentioned PC that Suse Linux 10 operating systems have been installed with the root identity, editor's share directory configuration file/etc/exports specifies share directory and authority.Carry out as issuing orders:

/datapool/share?107.252.78.*(rw，sync，no_root_squash)

Rw: read/write authority, the parameter of read-only authority are ro;

2), carry out as issuing orders its enable port mapping:

#/etc/rc.d/init.d/portmap?start；

#/etc/rc.d/init.d/nfs?start。

The user also can restart the Linux server, starts nfs server automatically.After nfs server starts, also need to check the settings such as fire wall (generally need close firewall services) of Linux server, guarantee not mask NFS port that uses and the main frame that allows to communicate by letter, mainly be to check Linux server iptables, the setting of options such as ipchains, and/etc/hosts.deny ,/etc/hosts.allow file.

Through above-mentioned configuration, at first on the Linux server, carry out the loopback test of nfs server, whether the checking share directory can be accessed.Operation is as ordering on the Linux server:

#mount-t nfs 107.252.78.165 (this machine IP) :/datapool/share/mnt

#ls/mnt

Mentioned order is mounted to the NFS of Linux server output share directory/the mnt catalogue under, therefore, if the NFS operate as normal should be able to be seen in/mnt catalogue/ content in the datapool/share share directory.

Step 3: utilize each operating system access network file system(NFS) storer.

Described each operating system is in (SuSE) Linux OS, AIX operating system, HP-UX operating system and the solaris operating system at least, wherein:

#mount-t?nfs?IP:/share/test；

#nfso-o?nfs_use_reserved_ports＝1；

#mount?IP:/share/test；

#mount-F?nfs?IP:/share/test；

#mount-f?nfs-o?vers＝3?IP:/share/test。

As shown in Figure 3, Fig. 3 is the structural representation that utilizes the witched-capacity network file system memory that the present invention builds.In Fig. 3, the built-in disk of four 1T is connected on the PC that has 4 SATA II interfaces, Suse Linux 10 operating systems have been installed on PC, and PC has been connected in the internal network, the server of large capacity disc being shared to each operating system by the mode of NFS uses, and reaches the purpose that various operating systems are used the high capacity storage temporarily.

Below the witched-capacity network file system memory of putting up is tested:

In the PC-SATA-NFS storage of putting up, we have carried out data transmission test, have obtained following data.

	The SAN storage	The NAS storage	PC-SATA-NFS	Built-in disk
					This accessing (reading)	-	-	126.0MB/s	58.2MB/s
This accessing (writing)	-	-	70.5MB/s	50.3MB/s
					Remote access (reading)	77.2MB/s	54.8MB/s	41.7MB/s	-
Remote access (writing)	73.3MB/s	46.9MB/s	35.3MB/s	-
					The storage price	85000 yuan/TB	160000 yuan/TB	600 yuan/TB	30000 yuan/TB

In the test of this accessing, the PC-SATA R/W speed of hard disc has surpassed minicomputer built-in disk read or write speed.

In the remote access test, SAN storage read-write speed is the fastest, secondly is NAS storage and PC-SATA-NFS storage.But PC-SATA-NFS and NAS storage speed are more approaching.

On the storage price, the storage price of PC-SATA-NFS is the most cheap, and it is more many than the low price of other three kinds of storages.

Above-described specific embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any modification of being made, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims

1. a method of building the witched-capacity network file system memory is characterized in that, this method comprises:

Step 2: personal computer is configured to the network file system(NFS) storer.

2. method of building the witched-capacity network file system memory according to claim 1 is characterized in that, described step 1 specifically comprises:

Select for use the conduct of Suse Linux 10 operating systems to set for this platform, and use the personal computer that has a plurality of SATA II hard-disk interfaces.

3. method of building the witched-capacity network file system memory according to claim 2, it is characterized in that, described Suse Linux 10 operating systems carry network file system(NFS) management server function, and this operating system carries software and has the Redundant Array of Independent Disks (RAID) function, can be according to the different demands for security of storage data are selected for use suitable Redundant Array of Independent Disks (RAID) grade.

4. method of building the witched-capacity network file system memory according to claim 2 is characterized in that, described step 2 specifically comprises:

Land the personal computer that Suse Linux 10 operating systems are installed with the root identity, editor's share directory configuration file/etc/exports specifies share directory and authority.

5. method of building the witched-capacity network file system memory according to claim 4 is characterized in that, described editor's share directory configuration file/etc/exports specifies share directory and authority, by carrying out with the realization of issuing orders:

/datapool/share?107.252.78.*(rw，sync，no_root_squash)

Rw: read/write authority, the parameter of read-only authority are ro;

2), carry out as issuing orders its enable port mapping:

#/etc/rc.d/init.d/portmap?start；

#/etc/rc.d/init.d/nfs?start。

6. method of building the witched-capacity network file system memory according to claim 5 is characterized in that, starts network file system server in the described step 3) and also can adopt the mode that restarts the Linux server to realize.

7. method of building the witched-capacity network file system memory according to claim 5, it is characterized in that, after network file system server starts, also need to check the fire wall setting of Linux server, close firewall services, guarantee not mask network file system(NFS) port that uses and the main frame that allows to communicate by letter, check the setting of option iptables, ipchains in the Linux server, and inspection/etc/hosts.deny ,/etc/hosts.allow file.

8. method of building the witched-capacity network file system memory according to claim 7 is characterized in that, through after the above-mentioned configuration, this method also comprises:

#mount-t nfs 107.252.78.165 (this machine IP) :/datapool/share/mnt

#ls/mnt

9. method of building the witched-capacity network file system memory according to claim 1 is characterized in that, this method also comprises:

Step 3: utilize each operating system access network file system(NFS) storer.

10. method of building the witched-capacity network file system memory according to claim 9, it is characterized in that, described each operating system is in (SuSE) Linux OS, AIX operating system, HP-UX operating system and the solaris operating system at least, wherein: