IBM Copy Service -- Flashcopy Introduction

IBM provides three advanced features of copy service
1. flashcopy
2. volumecopy
3. Enhanced remote partitioning ing (ERM)

This article describes flashcopy, volumecopy, and ERM.
1) First, let's take a look at the description of flashcopy in redbook.

A FlashCopy is a virtual logical drive that is a point-in-time (PiT) image of a real logical drive. the FlashCopy is the logical equivalent of its complete physical source, at the time created, but it requires less physical disk space and is created quickly when compared to the time required to create a full copy of the source. flashcopy is a real lun PiT image, which requires only a small amount of physical disk space. Compared with creating a complete copy, flashcopy is quickly created because Repository is only used to store changes since it was created ).

The real logical drive that is copied is known as the base logical drive. you can create up to four FlashCopies from each base logical drive. when you create a FlashCopy, the controller suspends write operations to the base logical drive for a few seconds as it creates a FlashCopy repository logical drive. the repository is a physical logical drive where FlashCopy metadata and copy-on-write data ar E stored. When a flashcopy is created, the Controller suspends write operations on the source volume for several seconds because it creates a flashcopy repository lun. This repository is the physical LUN that stores flashcopy metadata and copy-on-write data]

Using the FlashCopy feature requires less disk space, because the repository logical drive is used for storing only data change operations that have been made since its creation. because the FlashCopy is not a full physical copy, if the source logical drive is damaged, the FlashCopy logical drive itself cannot be used for recovery. flashcopy only occupies a small amount of disk space. Because flashcopy is not a complete physical copy, if the source volume is damaged, flashcopy cannot be used for restoration]

2) Flashcopy can be used for the following purposes:

---- Deming backup images

---- Creating a temporary test image

---- Creating a source image for a VolumeCopy operation that can be copied without extended effects to the production operations.

 

3) The specific principle is as follows. As mentioned above, the repository lun only stores some metadata pointer map and copy-on-wirte data. Because it is not the real copy data, the creation speed is very fast.

650) this. width = 650; "style =" border-right-0px; border-top-width: 0px; border-bottom-width: 0px; border-left-width: 0px "title =" Image (3) "border =" 0 "alt =" Image (3) "src =" http://www.bkjia.com/uploads/allimg/131228/0336063405-0.png "width =" 644 "height =" 428 "/>

When a flashcopy is created, the Controller suspends write operations on the source volume for several seconds because it is creating a flashcopy repository lun. This repository is the physical LUN that stores flashcopy metadata and copy-on-write data. From this we can guess that if the data of the base lun changes frequently, the repository lun will grow faster. The default size of the flashcopy is 20% of the base lun, however, this value can be adjusted or adjusted after flashcopy is created. But what if the repository is full but the Administrator does not know? There are two policies:
1. The base Lun continues to write, and the flashcopy lun does not have a valid. This is the Default policy, which ensures that the application continues to write data.
2. The base lun cannot be written, and the flashcopy lun is still valid because no new copy-on-write data is generated. To ensure that the repository lun can be used for restoration, the Host application is prohibited from continuing to write data to the base lun.

4) using defragmentation on the base lun will result in copy-on-write of each data block of the base lun. This will cause the flashcopy repository lun to be full when defragmentation is complete. To prevent this, ensure that the flashcopy repository lun is 105% of the base lun during defragmentation of the base lun. This is the minimum value of copy-on-write for each data block of the base lun.

Note: defragmentation is literally called "fragment", but this is not part of the file system that we often hear about, the specific content will be described in ibm ds Storage Other function.

 

5). recreate flashcopy

If you do not need flashcopy, you can disable it and do not have to delete it. If you want to create flashcopy for the same base lun on which day, you can use re-create

Recreate flashcopy deletes all copy-on-write data on the original repository lun.

Recreating flashcopyInvalidatesThe current FlashCopy and creates a new point-in time copy.

 

PS: flashcopy is easy to do. For detailed operations and steps for restoration, see the next IBM copy service-Flashcopy experiment.

This article is from the "star & storage" blog, please be sure to keep this source http://taotao1240.blog.51cto.com/731446/1002206