Case:
HP EVA4400, 5*12 blocks 1T physical disk, all storage spaces form two Storage Group, allocating Lun about LUN: 9TB 1, 8TB of about 4, LUN mapped to two Windows Server.
Solution:
Back up the fault storage in a read -only way, and then no longer operates the original storage. The specific implementation rules are about: divide 5 LUNs used to place mirrors on the new P6300 storage. In order to be obviously different from the source storage, to prevent misunderstanding, the size of these 5 LUN is slightly greater than the fault LUN. The fault storage EVA4400 and mirror target storage P6300 received a HP DL380 server at the same time. Install the software Winhex, a third -party publicly distributed software on the Windows system on DL380. Use Winhex to store only mirrors for source faults. Engineers with more than 2 or more mirror processes confirm the operation. After each group of LUN backup is completed, the source fault lun map can be removed to avoid misunderstanding during recovery. The logical analysis is performed in the backup carrier, and the general process of analyzing is: analyze whether the GPT format is abnormal, such as abnormalities, calculate the reconstruction. Analyze whether the FILESYSTEM DBR is abnormal. If there are abnormalities, reconstruction is calculated. Analysis of whether the NTFS $ MFT FileRecord structure is abnormal. If there are abnormalities, re -generate the correction through the NTFS $ logfile or upper and lower verification. Analyze whether the NTFS INDX is abnormal. If there are abnormalities, re -generate the revision through the NTFS $ logfile or context. After confirming in the virtual environment, loading the file system in the backup environment, CHKDSK can see if it can be verified by the system. If the reasons are not analyzed, repeat the above process, or make a response based on the situation. Data from the fixed LUN to process the next LUN. After all the data is completed, verify and transfer to complete the business process.