Today we want to write about another successful data recovery case from a damaged RAID array. The drive failure of the server, causing data loss, severely impacted the operations of the business. Yesterday, we successfully recovered the business file server from a failed RAID 5 array configured on the HP ProLiant ML380 G6 server.
The cause of the failure of the logical RAID server volume was a mechanical failure to 2 hard drives. The RAID volume was configured using three hard disk drives. The customer wanted us to recover data stored on the D partition of the HP ProLiant server.
The users of the organization used the central shared files on this partition. Therefore, there were lots of documents and various files formats.
HP ProLiant ML380 G6 Server Recovery
The model of the server was HP ProLiant ML380 G6. It is a reliable server if you configure and maintain it correctly. It is possible to extend its capabilities by installing lots of RAM, hard drives, and up to 2 Xeon Processors.
The RAID controller installed in this server was HP Smart Array P410i. It supports such RAID configurations as RAID 0, RAID 1, RAID 5, and RAID 50 array configuration.
The hard drives connected to the server were SATA ones. HP server also supports SAS drives, but the customer, for some reason, chose cheaper drives. Unfortunately, the SATA hard drives are less reliable than SAS hard drives.
Still, these hard drives took this company about five years till now. Overall, it is an excellent lifetime period for SATA hard disk drives.
Evaluation of RAID Array Members
All three hard drives were connected to the professional hard drive diagnostics tool and tested thoroughly. Unfortunately, all the hard drives were very noisy, the reading speed was an average of 5MB-10MB per second, and there were many bad sectors.
Seeing that the hard drives have physical issues, we could not run data analysis tasks to identify the RAID types and evaluate the chances of the data recovery process. The further we run heavy analysis tasks directly on the hard drives, the higher the chances of clicking and scraping the surfaces.
Bit-by-Bit Cloning of the Hard Drives
Thus, we decided to clone the hard drives to a binary image and continue to work with the binary files. This way, we prevented the risk of damaging source devices. For this case, the size of the hard drives was not significant, and it did not take lots of time to clone the hard drives.
However, the cloning process might take days, but the data recovery case was time-sensitive with high-capacity hard drives. The capacity of the hard disk drives was not the same. 2 of them were 160GB, and one was 500GB. We expected two hard drives with the same capacity to be either RAID 1 or RAID 0 array. But further analysis denied this expectation.
HP ProLiant RAID 5 Recovery
RAID 0 and RAID 1 configurations do not use parity blocks, but RAID 5 does. We focused on RAID 5, not RAID 6, because the controller does not support the RAID 6 array type.
The entropy analysis on each hard drive using the correct block size and drive count revealed positive results. We understood that these hard drives were configured as RAID 5 volume with 128 sector block size and 16 block delays.
Once you have a complete copy of each RAID member and precise RAID parameters, it is easy to set up the logical volume using any RAID server data recovery software. But there is one crucial thing to consider. If one of the RAID members stopped working a long time ago, using this hard drive to construct the logical volume will result in damaged files.
It is essential to work only RAID member drives with new data in them. Our engineers could construct the logical volume and copy all the customer data from the D partition to the USB external hard drive using the advanced RAID recovery software. There was no need to restore the server itself as the customer decided to use a new server with new SAS hard drives.