RAID 5 Set Suddenly NOT VIABLE!

Brett,

I have just experienced a similar problem to your earlier issues for the second time in 7 days - one of my drives (both times Bay 3) has dropped out of a Raid 5 set. I have an early 2009 MacPro with four 1TB drives. The first time I simply reassigned Bay 3 as "Make Spare" and the RAID set was rebuilt (i.e. as per your earlier experience). Twice in one week is too much to accept as a small issue and I fear I may be heading down the road to your failure described above. I see you've had no response and wonder whether you have any further thoughts. For myself, I am beginning to suspect the RAID card, as indeed you suggest.

I hope you didn't lose any of your data.

Thanks

Nick

In my experience, Apple RAID in general (with RAID card or without) does not tolerate a very high error-retry rate from the Drives.

I had a mirrored RAID that repeatedly became Degraded, and decided to take strong action. I made sure I had two Backups, then pulled out the Drives and erased them with Write Zeroes, one pass. One of the drives had more than 10 blocks that needed to be "spared", so Initialization Failed. I ran the procedure multiple times and found that 30 blocks had needed to be spared. If any of these had been marginal in daily use, it could have caused the drive to get (what appears to be) stuck doing dozens to hundreds of retrys, but still not throw an I/O Error that would bring the system to a Halt.

Google did a very large study and concluded that there is a very high probability that drives with multiple errors will end up being replaced within about six months. They attributed this to a cascade of additional errors that often followed the first group detected, leading to the drive being declared unusable.

-------

In my opinion, just reseating the drive (combined with rebuilding, which does re-write the data blocks) may not be strong enough medicine to fix a recurring problem. A drive may repeatedly be doing multiple retrys on some data blocks, or may have Bad Blocks.

I suggest you Zero the troublesome drives (which takes many hours each). If Disk Utility finds more than 10 blocks that need to be spared, or if the drive runs out of available spares, it will report "Initialization Failed!" Consider such a drive not good enough for use in a RAID and replace it with a new one. Or just rebuild your RAID with different (e.g., larger) new drives.

Thank you Grant, that is most helpful.

Nick