RAID5 is evil

[blu at vl.com (Tom Metro)]

Matthew Gillen wrote:
> I guess you didn't follow some of the links from Monday about why RAID5 is
> evil ;-)
> http://www.miracleas.com/BAARF/RAID5_versus_RAID10.txt

Nope, but I've posted in the past about my distaste for higher-order 
RAID and LVM due to the way they complicate recovery.

Practically speaking, it comes down to the lesser of several evils. If 
cost wasn't an issue, I'd opt for RAID1. RAID5 is the next best 
compromise to gain some redundancy while not eating up as much usable space.

The article you quote seems a bit dated, however (just consider the 
reference to $1000 disk drives, for example):

   Now SCSI controllers reserve several hundred disk blocks to be
   remapped to replace fading sectors with unused ones, but if the
   drive is going these will not last very long and will run out and
   SCSI does NOT report correctable errors back to the OS!  Therefore
   you will not know the drive is becoming unstable until it is too
   late and there are no more replacement sectors and the drive begins
   to return garbage.

   [Note that the recently popular IDE/ATA drives do not (TMK) include
   bad sector remapping in their hardware so garbage is returned that
   much sooner.]

Modern IDE/SATA drives that support SMART[1] monitoring will (given you 
are running the necessary daemon) notify you when they are starting to 
fail, and as implied by the Wikipedia article, such drives also 
automatically reallocate sectors.

1. 
http://en.wikipedia.org/wiki/Self-Monitoring%2C_Analysis%2C_and_Reporting_Technology


   When a drive returns garbage, since RAID5 does not EVER
   check parity on read...when you write the garbage sector back
   garbage parity will be calculated and your RAID5 integrity is lost!

This seems to ignore the integrity checks performed by the file system 
layer. Is it not typical to at minimum have sector checksums in all 
modern file systems? The RAID layer may not be aware of the failures, 
but your OS will be.


   What about that thing about losing a second drive?  Well with RAID10
   there is no danger unless the one mirror that is recovering also
   fails and that's 80% or more less likely than that any other drive
   in a RAID5 array will fail!

As another article recently posted to the list points out, that 80% 
probability isn't quite right. In practice the probability of a second 
drive failure is higher due to the process of recovering stressing the 
remaining drive. This is because the rebuild requires reading 100% of 
the data on the remaining drive, so any weak sectors will be uncovered. 
The article noted that performing regular backups will similarly stress 
the drives, and catch the problem before it becomes critical.


   The original reason for the RAID2-5 specs was that the high cost of
   disks was making RAID1, mirroring, impractical. That is no longer
   the case!

Of course all things are relative. If you're storing a modest amount of 
data and you have a corporate budget, sure, mirroring is plenty cheap. 
If you're trying to store 1 TB on a home server, the cost of mirroring 
is prohibitive. Besides, RAID should only be used for increased 
reliability (decreased down time), not data integrity, which still 
depends on traditional backups.

  -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.