We now have drives that are being manufactured with huge data capacities for literally pennies per gigabyte.
We can now easily configure 5+ Terabytes of RAID 5 or RAID 6 storage arrays with ease for those clients whose data volume needs require huge storage capabilities in a very inexpensive manner.
Ever have to selectively recover 600GB or more of data off of a large failed RAID 5 array because someone "forgot" to change the tape magazines ... for quite a while?
Or, come into a situation where a business critical workstation has had a drive failure with no RAID 1 mirroring or any kind of backups being done?
Or, a hard drive has failed in a RAID 5 or RAID 1 array, and the second one tanks before the rebuild completes, thus rendering the array useless? Again, backups were sporadic at best.
The general theme here is this: Poor or sporadic backup schedules are the death of a business in the event of (not if) a storage failure.
With existing hard drive technology where the storage bits lie side-by-side, all of the effort a data recovery may take is for the drive to be stored overnight in a static bag enclosed in a freezer bag (air vacuumed out) and placed in a deep freeze to render the drive useful enough to recover the necessary data the following morning. We have a freezer here in the shop just for that purpose.
Or, in the case of a failed on-board drive controller board ... swapping that controller board out with a known good one to get to the data for recovery. The defective drive would still be replaced by a new one.
If we are dealing with sectors of a drive failing, with the current single bit technology drives we may have a fighting chance of recovering data from the bad sectors using the above freezer method, a really good software recovery tool like Get Data Back for NTFS, and a powerful enough data mule system to slave that drive up to. Of course, one needs a lot of patience in some cases waiting for things to come up or the recovery to complete.
With the advent of Perpendicular Magnetic Recording (PMR) technologies for hard drives, we need to take very careful stock of our client's data storage and backup setups.
PMR places five (5) bits stacked on top of each other for storing data side-by-side.
Consider PMR again: We have gone from one bit sitting beside another bit and so on to five bits sitting beside five bits and so on.
What does that mean for a bad sector scenario? Instead of a few bits being obscured by a sector going bad, we are now talking about chunks of bits being lost.
The other consideration that we need to keep in mind: When we defragment our server and desktop drives now, the algorithms being used to reorganize files on the drive are based on the bits being side-by-side.
We will need to pay particular attention to the manufacturer of the utilities that we use to defragment our drives with and whether there needs to be an upgrade to that utility to be compatible with the PMR based drives.
Please keep in mind that we do not necessarily know the deep down engineering of the new PMR technology and deframenting them ... so this may be a non-issue. But, the technology is too new to know for sure yet.
Our experience does tell us however, that the more prominent that PMR drives figure in our system configurations, the more we will need to keep in mind the possibility of data being non-recoverable from a failed PMR drive.
- Seagate's Press Release for 1GB Drives: Seagate Unveils One Terabyte Hard Drives as Explosive Growth of Digital Content Continues.
- Seagate ST3100034NS 1TB Barracuda ES.2 SATA Drive
Microsoft Small Business Specialists
*All Mac on SBS posts are posted on our in-house iMac via the Safari Web browser.