Is it possible/interesting to reconsider a RAID array online?
ZDNET placed a really debatable write-up concerning RAID5 doomed in 2009 as a result of larger and also larger abilities time ago (2007 ).
2009 has actually gotten here and also I would love to recognize if it is feasible to reconsider a RAID array online : read the entire discs and also at some point identify a disk failing beforehand (ie exempt to any kind of SMART or information - required basis).
I can consider a really straightforward means of examining : do a
dd if=/dev/md0 of=/dev/null and also check the RAID logs, yet I'm looking of an extra incorporated means, that can also do some magic with the mistakes (like transferring them).
Linux software program RAID (md) in several circulations (debian & ubuntu at the very least) will certainly usually execute a resync month-to-month (first sunday) simply in instance ; seek the "checkarray" manuscript.
NetApp filer is (by default) do a reduced top priority resync every sunday early morning.
In technique I do not think I've ever before had either of these catch a trouble despite having thousands of systems in manufacturing.
Most modern-day array controllers give this as the controller degree. Dell PERC raid controllers have actually something called "Patrol Read" which works out the disks in the array.
ZFS has actually an attribute called "resilvering" which does something comparable.