IBM_V7000 infrastructure and server data recovery case Detailed

Detailed principles [IBM]
IBM_V7000 (full name IBM Storwize V7000) is IBM's next-generation midrange storage system, although positioning in the end, Storwize V7000 has provided in the past only high-end storage with powerful storage management capabilities. The common models IBM Storwize V3700, IBM Storwize V5000 and IBM Storwize V7000. Storing the underlying structural support RIAD 0, RAID 10, RAID5 and RAID 6. The upper roll support normal volume, compact mode volume, volume mirroring mode and the streamlining of the mirror mode volume. Although the V7000 to do very well in the overall storage architecture, but some failure or other physical operations or storage volume might cause damage, so the data stored on the series of recovery techniques have useless.
[Configuration IBM_V7000]
1, using the management IP connection IBM_V7000, enter your user name (default: superuser) and password (password: passw0rd).

2, there is no default configuration, you need to configure Mdisk, pool, and volume, then add host mapping volume.

3. Create Mdisk, type Mdisk are RAID0, RAID0, RAID5 and RAID6.

4, the pool is created, multiple Mdisk can be divided into a pool.

5, create a volume on the basis of volume is arranged above the pool, the volume type universal, thin, streamlined mirror, and a mirror.

6, and creates a mapping host volume, the type of a host in a host and a Fiber Channel host iscsi.

At this point the entire configuration of the general process even finished, but we do not know the logical volume allocated to the host, the actual disk in how they are distributed. That is how they distributed it? Detailed below is the structure and principles of its structure and HP Lefthand series of storage products like,
[principle] structure and
in fact the underlying principle IBM_V7000 is not complicated, the entire storage structure divided into a total of four.
First layer: both physical hard disk, is where the actual storage of data.
The second layer: IBM_V7000 named as Mdisk, in fact, RAID, is a collection of multiple physical disks.
Third layer: pool, is combined into a large plurality Mdisk logical container.
Fourth layer: volume dispensed from the pool space, facing the user of the storage unit, can not cross cell volume.
Structure is as follows:

the whole storage structure, local disk data is ultimately stored. The so-called Mdisk, pools and volumes are the physical disk virtualized only. This layer is the physical disk, the data is in units of small blocks (Block) storage, N Mdisk a plurality of disks, the data existing in the Mdisk N Block will be divided into a plurality of evenly distributed across all disks. Mdisk In this layer, the data segment is (Section) storage units, composed of a plurality of Mdisk pool, both are divided into a number of segments into different Mdisk in volume created in the pool, the volume of different types of distributed in the pool are also different, but ultimately the unit of segment stored in Midsk in.
  The whole procedure is stored user data will be stored to the volume, the volume will divided into several segments in different Mdisk, whereas Mdisk segments will turn into a number of blocks distributed in different disks. All final data are distributed in different units of blocks of disk.
[Data Recovery] Case
1, the storage architecture
storage model: IBM_V7000
number of disks: 24 600G SAS disk
number Mdisk: 2 Ge Mdisk, are RAID 5
Volume Number: 2 common pattern of 2T, 3T a compact mode.
2, the cause of the malfunction  
  due to aging cause Mdisk disk has a few disk dropped, resulting in unusable Mdisk cause upper volume inaccessible. And because only set up a global hot spare, disk management staff did not promptly replaced after dropping the disk that has caused the occurrence of the entire failure.
3, the solution
to all disk mirroring, disk and then analyze which is a set of Mdisk. Find belonging to the same group Mdiskd disk, and then dropped the disk situation analysis exists. If you dropped the disk exists, to get rid of this disk in Mdisk group. Generating Mdisk, followed by analysis of the structure between Mdisk. Generating a pool, followed by analysis of structure of a volume, because of different volumes of different types, the storage structure is not the same. After analyzing the structure of a volume after it may generate the data for each volume.
4, data recovery results are
  due only Mdisk there are a few disk is down, do not have other operations. So the entire data recovery is complete, user acceptance data is correct.
  Because of the storage infrastructure IBM V7000 series of very thorough research, so this series storage failure, data can be saved almost. However, there is a case where data can not be salvaged, and that is all disks are rebuilt and initialization is complete. Because after you've created Mdisk, the system will Mdisk do initialization, which is cleared. FIG follows

Thus, if the storage has been rebuilt, and also initializes, to recover data in this case it is desirable slim.