On 14-10-31 11:39 PM, Alex Pilon wrote: > On Mon, Oct 27, 2014 at 05:39:29AM -0400, Rick Leir wrote: >> On 26/10/2014 12:00 PM, linux-request [ at ] lists [ dot ] oclug [ dot ] on [ dot ] ca wrote: >>> ?: /dev/sdd: >>> reading sector 0: SG_IO: bad/missing sense data, sb[]: 70 00 03 00 00 00 00 0a 00 00 00 00 11 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 >>> succeeded >>> 5355 4342 002c 0000 0200 0000 0080 8510 >> Sense data are the markers that indicate the start of the next block, so at >> least part of the disk is corrupted at a very low level. Or the read head is >> non-functional. >> >> Did you try dd skip=BLOCKS to jump past block zero and try another track? > Yes. That and reading arbitrary sectors with hdparm --read-sector. No > change. Already falling back to the data recovery service. I'll let > this list know anything useful about their services. I hope you have success getting your data back. If you are referring to Tunstall & Tunstall, I've used them in the past - and highly recommended them to clients who have as well. They are pros - used by the RCMP and other government security agencies needing evidence recovery and skill/tool sets far above the average sysadmin. The cost of data recovery and the risk of relying on cheap drives taught me to build multiple redundant servers - I operate with at least two complete "hot spare" backups capable of being dropped into my co-location site and can be back up within the time it takes to drive the spare to the live site. I rotate boxes out of service after three years and generally have at least one new box every year in the cycle. As soon as I have any suspicion of a drive failure impending I promote one of the hotspares into production and build a new one. Backup servers are a lot cheaper than data recovery. -- Bill Strosberg