[ldm-users] raid / Re: Fedora 7 redux: BUSTED!

[Gerry Creager]

Gilbert Sebenste wrote:
> On Tue, 16 Oct 2007, Rob Cermak wrote:
> 
>> RAID 1?  I like Gerry's idea of moving from RAID 5 to 6.
> 
>> We had a 750GB disk die.  The volume affected went in to degraded mode.
>> We swapped the disk out and the volume regenerated automatically.  Took
>> 4.5 hours, but it regenerated.  During this 4.5 hour rebuild, two things
>> can happen:  (1) you can lose another disk and your whole volume is
>> toast, (2) find out the new disk you swapped in is bad.  Anyway, this
>> whole proceedure did not require any downtime.  The OS (Mandriva 2007)
>> didn't notice a thing.  Granted I/O was a bit slower during the degraded
>> period and rebuild.
> 
> Still, I can accomplish the same (pretty much) and within budget with
> RAID 1. If I lose a disk, I can still replace it with no downtime. That's 
> the most important thing I need. And within my budget.

And that is a serious consideration.  We have to store in perpetuity, 
and it gets messy if we start losing data.  Been there once, had one 
helluvatime getting data restored.  We lost one disk and initiated the 
rebuild (automatically) but lost two more drives almost immediately. 
Unrecoverable.  Good news: we had a distributed archive site with the 
data and we were, over time, able to repopulate the 2 or so TB we had 
lost.  Our database told us what should be recovered and a subsequent 
audit showed we had recovered everything...

But, go with what meets your needs and fits your budget.  If you, like 
me, are in academia, sometimes meeting budget is more important than a 
lot of other considerations.

gerry
-- 
Gerry Creager -- gerry.creager at tamu.edu
Texas Mesonet -- AATLT, Texas A&M University
Cell: 979.229.5301 Office: 979.862.3982 FAX: 979.862.3983
Office: 1700 Research Parkway Ste 160, TAMU, College Station, TX 77843