Data failure scenarios brief:
Users using two disk composed of a raid0, where a bright yellow plate, after being kicked out raid card, raid crash, focusing on the following account of what the situation was rescued data.
A hard disk used by the user two SAS 300G, hard disk removed from the machine first, and then by way of SAS HBA win is directly connected to the environment, and the hard disk management flag in the offline state, to ensure that during operation is read-only, secure the original data.
After two hard bottom After all of the sectors are mirrored out by the file system analysis disk stripe size and sequence, using software virtual recombination in raid take up the original environment, and then further resolved finally see ntfs file system data , new problems emerged at this time, after so directly copy the data out of the original system and applications need to be deployed, but also because there is no software support service providers, to implement not easy to achieve, so thinking about the ride up the raid and then complete migration to the new raid environment before damage can be done and the same. Save a lot of time.
For new customers learned a lesson raid environment, we decided to use three disk set up raid5, so even if a disk problem goes offline, you can also raid in a degraded state, will not collapse immediately, giving the user the opportunity to replace a new hard disk.
After the installation of new support raid5 raid card and insert a new hard disk to create a volume raid5 now began to study how to do data migration out of the past.
Data migration overview:
Because the server front panel is managed by the raid card, so directly insert a new disc up is not directly identified in the system needs to be created in the raid after raid card can be used, but limited to a single disk capacity problem, this program can not be used, then study other way.
Since the front panel of the server has a DVD drive, and CD-ROM and the server now are based on the motherboard sata channel connection, the cover can be opened using the above sata port connected to a hard disk sata, pe at or linux live cd mode can move back up data, and this is the fastest method. But I found that this is not a standard size sata interface type of machine to use, but mini sata, because there is no ready-made adapter card in hand when we are ready to implement, so this approach does not work, in fact, the amount of data, they also can using a USB way to do it, but because the vast majority of servers now usb still in USB2.0 speed is too slow for the large amount of data, an unacceptable time.
Finally, the last we absolutely use a novel way to go fetch data ---- network.
This time we need to start a linux live cd Typically we use linux system rescue cd, boot to complete the latter ifconfig configure the server ip in linux, then we will do out of the data on a win 2008 r2 machines in the win environment under open nfs service (default is off) "service Manager - role - add roles - check file service - check for network file system installed, need to restart after the first installation at the computer" 
reboot is complete after we file folder to store image data of the operation carried out, right -NFS share tab inside check share this folder, and then there is a focus on the rights which check allow root access, read and write access type selected. 
Win after the end of the setup is complete, we look at the next set linux side, ifconfig to view the current network configuration. 
Because we need to allocate a ip to him, here we assign card "enp4s0", ip address assigned to 10.3.12.3 subnet mask of 255.0.0.0, use the following command: ifconfig enp4s0 10.3.12.3 255.0.0.0 and then use ifconfig Check ip address. 
Once you've configured ip, check whether the network along with the command: ping 10.1.1.1,
Whether to view the NFS shared directory on the 10.1.1.1 machine can access the command: showmount -e 10.1.1.1. 
The source machine and the target machine is now connected through, create a directory mkdir / mnt / bysjhf in linux
Once you've created, we will mirror out of the data under linux mount the newly created folder mount 10.1.1.1:/data / mnt / bysjhf -o nolock
after mounting well, look at the mount point information df -k 
determined after it has been mounted well into this folder, look at the folder in the image file:
root@sysresccd /mnt/bysjhf % ls
And view the hard disk and partition information: fdisk -l 
after confirming good source and target devices, mirroring operation:
dd if = / mnt / bysjhf / data.img of = / dev / sda bs = 10M 
in a Gigabit Ethernet environment, the speed can go NFS 70M / S, this is a very desirable rate, after the completion of the waiting dd we restart ibm X3650 server, and select raid guide, look forward to the start page of the windows finally appeared in front of the hard work paid off, with complete data migration success.