How to Recover the Lost Configuration Data of the Self-developed Storage System?

Publication Date:  2012-07-22 Views:  310 Downloads:  0
Issue Description
A Huawei research institute managed the S2300 by using the USM software (developed by Huawei Application and Software product line), but the DB data of the S2300 was lost by misoperation. As a result, 3 TB mission-critical data failed to be employed and storage data was on the edge of lost.
Alarm Information
None
Handling Process
Step 1  Forbid on-site engineers to perform further operations.
Step 2  Collect the storage information. Query and analyze the information about RAID groups and LUNs in the runlog file. The desired RAID group information includes the information about member disks, RAID levels, RAID group names, and RAID group IDs. The desired LUN information includes the LUN sizes, stripe depths, LUN names, and creation sequence. (The keyword to be queried is add.)
Step 3  The version of the storage system onsite is 1.02.01.204.T02, which is an early version and does not support the function of creating LUNs without formatting them. Therefore, it is recommended to upgrade the storage system to 1.02.01.222.T04 (V100R001C02SPC024) and SES version to S606.
Step 4  Enter the CLI and run commands to restore the previous RAID groups and LUNs, and not to format LUNs after their creation.
Command format:
  1. To create a RAID group: creatrg –n RAID name –i RAID level –d member disk slot (such as 0,1; 0,2; and 0,3).
If the RAID group to be created is a RAID 10 group, the –m parameter is added after the –i parameter, indicating the number of member disks in each RAID 1 sub-group. The number is usually an even number, and that recorded in logs prevails.
  1. To create a LUN: createlun –i RAID ID –n LUN name –s LUN size (unit: MB) –u stripe depth –f not to format the LUN after its creation.
Note: If a LUN larger than 2 TB is created by using the controller software of an early version (just as the version involved in this example), the actually created LUN is 512 MB smaller than the entered value (the problem is solved in the latest version). Therefore, the log information may indicate that the capacity of the RAID group is smaller than the total capacity of its LUNs. You can reduce 512 MB from a 2 TB LUN before creating it, and then map the LUN to the host.
Root Cause
Huawei Application and Software product line has developed a USM software specific for the management of storage systems (similar to the unified management software for blade servers). The USM provides the functions for easy configuration and modification of the storage system information. Furthermore, the USM owns a function to clear DB data. The research institution in this example triggered this function by mistake, thus deleting all the DB data and causing risk to the 3 TB mission-critical data.
Suggestions
This method is used to recover the DB data that was not backed up but lost, and is still of risk of data loss. It is recommended to back up the DB data beforehand. It is recommended that Huawei delete the function to clear DB data from the USM software to avoid hidden risks.

END