storage myths busted
Well, for storage freak^H^H^H^H^Henthusiasts it’s been quite a week. The only thing missing is the Mythbusters themselves presenting the facts. I guess I’ve had a fair share of failing disks (that link is just one of them) and once you’ve experienced this random behaviour on nowadays tech, anyone who likes to be in control of his or her tech becomes a storage fanatic just like that!
click to go to the corresponding pic on flickr
Anyway, first of all Google put a paper online with some interesting facts.
Failure Trends in a Large Disk Drive Population
link to pdf-file
There’s a summary to be found over here.
And then another paper became popular:
Disk failures in the real world:
What does an MTTF of 1,000,000 hours mean to you?
link to pdf-file
A summary of that paper can be found over here.
Read both papers (or their summaries) and really, you will think different about hard drives from now on (even when you didn’t think about them at all before)
For instance (out of the last paper):
“Interestingly, we observe little difference in replacement rates between SCSI, FC and SATA drives, potentially an indication that disk-independent factors, such as operating conditions, affect replacement rates more than component specific factors.ā€¯
By the way, both of these papers were presented on last week’s 5th USENIX Conference on File and Storage Technologies (aka FAST’07)
So do like Google does, no RAID and replicate everything three times (on cheap drives)... If this means anyone has to get rid of all-of-a-sudden-out-of-fashion SCSI U320 drives, I’ll be happy to oblige and buy some of those drives cheap. (I’m such a nice guy and I don’t mind some old-fashioned setups) *^_^*
permalink - 
add to del.icio.us
previous: ...and then, nominated for a Dutch Bloggie?!?
next: project: LED throwie bike lighting
I’M looking in to ZFS (the new FS from sun)... At least the darn thing notices when files are corrupted and adding more space to a pool is easy.