View All R&D Articles

Comprehensive RAID Array Recovery, Part 2

June 24, 2014

Part two.  Be sure to read part one of our comprehensive RAID array data recovery guide.

Step Four: Determining the RAID Algorithm

If the RAID array has a distributed parity, our engineers must determine the RAID’s XOR algorithm and, in some cases, the order of the member drives. This can be a tedious process if the drive order is not apparent, but engineers can provide results regardless of the number of drives in the array.

Regardless of the operating system, RAID 5 uses one of four algorithms for disk placement. NOTE: If you are attempting to repair a damaged RAID array, we strongly recommend contacting our RAID experts at 1.800.237.4200; do not attempt to determine your RAID’s drive order if you do not have significant experience with RAID troubleshooting.

The RAID parity rotation variations include:

  • Left Asynchrous – In this layout, data segments are numbered sequentially starting with the first non-parity drive in the stripe. The parity starts at the last drive and moves backwards (one drive per stripe).This is the hardware standard RAID layout, but it is not default for either Windows 2000, 2003 Server, or Linux. The left asynchronous layout is sometimes referred to as “backwards parity” or “standard rotation.”
  • Left Synchrous – This layout uses sequentially numbered segments that start with the first drive in the stripe after the parity. The data segments wrap. The parity starts at the left-most drive, then moves right (one drive per stripe).Linux uses the left synchrous layout as a default.For large reads, left synchrous is the most efficient option. Each consecutive group of segments will use all of the disks in the array, provided that those segments are no longer than the total number of hard drives in the array.
  • Right Asynchronous – Also known as forward parity rotation, the right asynchronous layout uses sequentially numbered segments that start at the first non-parity drive in the stripe. The parity starts at the right-most drive and moves left (again, one drive per stripe).
  • Right Synchronous – Once again, segments are numbered sequentially starting with the first drive after the parity. The parity starts at the right-most drive and moves left with one drive per stripe, and the segments wrap.

Drive order is an extremely important part of the RAID data recovery process. Our engineers need to determine the parity rotation and order before they begin destriping the array, or the resulting data will be completely unusable. We use a variety of tools to determine parity rotation based on the configuration of each array.

Step Five: De-Striping the Array

By using software developed in our research lab, we can destripe the reconstructed RAID array fairly easily. Our tools copy stripes from each member of the array in a manner that corresponds to the initial RAID configuration. These tools can also apply XOR calculations.

When the process is complete, we have a complete copy of the data on your array as it existed prior to failure. However, there could still be some corruption and file system errors, especially if your array was operated after losing one or more members or if you carried out a rebuild attempt.

Step Six: Making File System Repairs

To return your files in a working condition, we need to repair the file system by locating errors and determining whether data slippage has affected the volume. Our engineers need to ensure that essential parts of the file system are present.

As an example, consider NTFS, a common RAID file system used primarily with Windows systems. Windows creates a master file table $MFT along with metadata files, and $MFT is a relational databases consisting of rows of file records and columns of file attributes. $MFT contains at least one entry for every file on a given NTFS volume – including an entry for the $MFT itself.

Small files sized at about 1.5 KB are stored completely within the $MFT. This allows our engineers to match information within the MFT without access to the actual file being referenced, and we can use a specialized program to verify the file system and to correct for any data slippage that occurred during the data recovery process.

After recovering data from a RAID array, we copy it to the media of your choice and return it. Highly essential files can be returned electronically in some cases. To learn more or to begin the RAID data recovery process, call us today at 1.800.237.4200.