Storage redundancy of Windows Server storage space

Source: Internet
Author: User

Redundancy of storage space

In the high-availability architecture of the storage controller server, there is a difference in the degree of redundancy that is achieved by using a different number of JBOD or different ways of storage space layout. If you cannot turn on storage box-level redundancy by using two storage bins, so you cannot implement JBOD storage-tier redundancies, you must use three JBOD to achieve JBOD enclosure-level redundancy. If the storage space uses a dual mirror layout, the storage space allows for a maximum of one hard disk failure, and if two hard disk failures occur at the same time, the storage will fail, and with a triple mirror layout, a maximum of two hard drive failures are allowed. In the first-generation SDS structure of Windows Server, the storage space redundancy matrix is summarized below.

is not supported is not supported

Redundancy type

3 JBOD Storage Box

4 JBOD storage box

simple

dual mirror

Triple Mirror

A storage box and a disk;

Two pieces of disk

Two disks

single parity parity

Double parity check

Not supported

A storage box and a disk;

Two pieces of disk

storage space Availability Details

at the JBOD level, three JBOD can provide JBOD enclosure-level redundancy, because the storage controller is aware of a hard disk that is not in the same jbod for columns when the data is striped, and when the hard disk is selected for a column at the bottom. After the data strip is stored, each piece of data is mirrored by an I/O image in two or three copies, each written to a column, and the data is eventually written to the hard disk by the column. In the storage space of the mirrored layout, the different joins contain two or three hard disks, which come from a jbod or a jbod of different jbod that will directly determine the storage space's level of redundancy.

< Span style= "color: #000000; font-size:14px;" > at hard disk level, the virtual disk is mirrored and raid1 similar to the principle of selecting disks for data storage, but with raid1/raid10 The maximum number of hard disk failures is different, RAID1/RAID10 can remain available in the event of half of the hard drive failure. The Storage spaces column does not divide the hard disk symmetrically into two fixed groups or three groups and then writes the individual copies of the data, the column selects the hard disk followed by the randomness and the average principle, and the random description column each time a copy of the data for the stripe is selected for the written hard disk to be changed, The average principle describes the probability that all hard drives are selected when the hard drive is selected, and that the more data bands are closer to the average. As an example of a dual mirrored virtual disk, copy a 100GB size file to storage space, convert the unit to KB 104857600KB, the default interleave is 256K, write 409,600 columns, which means to write 409600 to the hard disk, To produce 409600 pairs of random hard disk drives, the average number of hard drives per data write is an inevitable event. Therefore, the dual mirror virtual disk even contains multiple hard disks, each hard disk must have a copy of the data block of the 100GB file, a hard disk failure data is still intact, but there is a second hard disk failure, there must be a loss of data blocks resulting in incomplete files, so can only withstand a failure of a hard disk damage. In the same vein, the triple image of the vdisk will replicate a copy of the data into three copies written on three different hard drives, and the virtual disk can withstand the failure of two hard disk damage.

Next we enumerate the different JBOD the number and storage space layout matrix Relationships example to illustrate how storage space achieves the expected availability goals.

    • Two JBOD storage Bins

< Span style= "color: #000000; font-size:14px;" > system Default Design is two jbod jbod jbod does not support enclosure awareness, when selecting an associated hard disk for a column, the column selects two blocks (dual mirror) or three hard disks (triple mirror) from the same must be an inevitable event, such a jbod failure, data loss of all replicas results in incomplete data, 1

650) this.width=650; "title=" 303-figure 1.jpg "style=" Float:none; "src=" https://s2.51cto.com/wyfs02/M00/8E/FA/ Wkiom1jqnmkdpbckaadjjjojrt8984.jpg-wh_500x0-wm_3-wmp_4-s_3710048234.jpg "alt=" Wkiom1jqnmkdpbckaadjjjojrt8984.jpg-wh_50 "/>

  figure 1 two jbod data storage schematic

    • three a JBOD storage box

three xJBODsupports storage box awareness, so three xJBODis the recommended configuration to implement storage box redundancy, and to enable the storage poolIsenclosureawareparameters. Three xJBODWhen you enable storage box awareness, the Storage space column selects the associated hard drive, and you will always choose a differentJBODthe hard drive inside the storage box writes data. For double or triple mirroring, no matterJBODHow many hard disks there are, and no matter how much data is written, since any strip of data after two copies and three parts are not in the sameJBODthe hard drive is stored, so aJBODFault, the other twoJBODmust hold additional copies of all data to preserve the integrity of the data,2is shown.

650) this.width=650; "title=" 303-figure 2.jpg "style=" Float:none; "src=" https://s2.51cto.com/wyfs02/M02/8E/F8/ Wkiol1jqnmwijx4faadeul1o0pi251.jpg-wh_500x0-wm_3-wmp_4-s_1771054980.jpg "alt=" Wkiol1jqnmwijx4faadeul1o0pi251.jpg-wh_50 "/>

Figure 2 three x JBOD data storage schematic

In addition to the storage box aware of this parameter, the role of the column is very important, the column is like an organizer, organized data to the hard disk to write, the storage box-aware control column in the selection of the hard disk when the same JBOD the hard disk "mutex" role, so that a column associated with the hard disk must not come from the same jbod. Discusses the availability implementation of JBOD levels, and then discusses the availability of JBOD in combination with storage space layouts.

    • three JBOD plus dual mirror

If the storage space is a dual mirror layout, as shown in 3, the DISK1 in JBOD1 can be randomly combined with any of the 6 hard drives in jbod2,jbod3, instead of looking for the other hard drives in the JBOD1 to be combined into columns. Similarly, the DISK5 in the JBOD3 can be randomly combined with any of the 6 hard drives inside the JBOD1,JBOD2. Dual mirroring of virtual disks using three JBOD and enabling the Isenclosureaware parameter, you can guarantee that two copies of the data will not be written in the same JBOD, a JBOD fault, a copy of the data remaining, the copy stored in the other two normal JBOD, so the data is complete. But at this time, any one of the remaining two JBOD may retain the only remaining copy of all data, no more JBOD failures and any hard drive failures, because any single hard drive failure can result in data loss. So, three jbod (storage box sense) plus dual mirroring this storage space virtual disk can withstand the limit of a JBOD or a piece of hard disk damage failure.

650) this.width=650; "title=" 303-figure 3.jpg "style=" Float:none; "src=" https://s4.51cto.com/wyfs02/M00/8E/FA/ Wkiom1jqnmrqm3eqaaemmmrkdzi794.jpg-wh_500x0-wm_3-wmp_4-s_4044004771.jpg "alt=" Wkiom1jqnmrqm3eqaaemmmrkdzi794.jpg-wh_50 "/>

Figure 3 three x JBOD and dual mirror data storage schematic

    • Three JBOD plus triple mirror

< Span style= "color: #000000; font-family: ' Song-body '; font-size:14px;" > save

650) this.width=650; "title=" 303-figure 4.jpg "style=" Float:none; "src=" https://s1.51cto.com/wyfs02/M01/8E/FA/ Wkiom1jqnm6ddb1kaaejs5pvlg0482.jpg-wh_500x0-wm_3-wmp_4-s_3315929354.jpg "alt=" Wkiom1jqnm6ddb1kaaejs5pvlg0482.jpg-wh_50 "/>

Figure 4 three x JBOD and triple image data storage schematic


This article from "Huang Lijun Blog" blog, reproduced please contact the author!

Storage redundancy of Windows Server storage space

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.