On 2/7/2013 4:35 PM, John P. Rouillard wrote: > > In message <[email protected]>, > "Brian O'Neill" writes: >> On 2/7/2013 2:04 PM, John P. Rouillard wrote: >>> Our normal decomissioning (before releasing to employees or sending it >>> out for recyling) is to run a single cycle dban with a zero pass for >>> normal drives. For drives with sensitive (PII or contractually >>> obligated data) we return the drives to the supplier or do a more >>> extensive wipe of the data using multiple methods. >> >> Ironically I'm at a datacenter all day today, wiping systems. >> Unfortunately DBAN failed to work on a number of systems (they all have >> hardware RAID), so we're using live CDs and a perl script I whipped up >> that writes pseudo-random data directly to the drives. > > We use 3ware, areca and LSI hardware raid. The trick with those is to > turn the disks into jbods. IIRC on 3ware, you destroy the raid, set > the controler to export jbods and type J or j to turn your disks into > jbod disks. On LSI you do it by creating a raid 0 consisting of a > single disk. Don't know how we do it on areca's. > > But once you have jbod'ed them you can see all the disks in the array > and dban has worked fine. The nice part is that you get to wipe all of > the disks in parallel which speeds up the wipe process. Also some raid > controllers trim the disks to a size slightly smaller than the actual > drive size (so you can mix and match drives with the same nominal size > but slightly different real sizes). Some raids use that extra space > for metadata. By jbodding you get to clean that off as well.
Yeah, we tried JBOD'ing the disks, and it didn't work for some reason. There are a variety of HP and Dell controllers. And we weren't terribly concerned with the reserved space, just the filesystem data. _______________________________________________ bblisa mailing list [email protected] http://www.bblisa.org/mailman/listinfo/bblisa
