Re: iffen ou want I could look at the backup disk

Post #299,215 by Ashton 11/8/08 8:21:46 PM Reply	Re: iffen ou want I could look at the backup disk You mention 'dd'; since this needs a disk assembly able to perform mechanically. electronically OK -- I presume this means it is Not a mechanical (or fried onboard electronics) issue with the backup HD (?) I mean.. if that HD merely expired mechanically (or, say the +5V went to 25, etc. wiping the electronics board) -- mightn't there be a decent chance of successfully swapping the physical disks to a new identical drive? (Via some contrived plastic hood with small blower, a stone-air-filter of the sort used by watchmakers, for small positive air pressure inside -- as a disposable clean-room.) I realize, of course, that for big-$$ there are many places who do this routinely; I always prefer gaming the System, when I think I can / or won't spend the $$. Ashton
Post #299,216 by Andrew Grygus 11/8/08 8:51:49 PM Reply	If it's an electronic problem . . . . . it could be overcome by swapping the circuit board with an identical one. The problem is, it's almost impossible to find an identical board - seems the manufacturers change rev levels every couple of months or so, often changing geometry and head interfaces. Swapping disks into an identical drive is a job for specialists who charge thousands of dollars to do it.
Post #299,223 by Ashton 11/8/08 11:51:46 PM Reply	Ah, about the first - bummer. So then, only the disk-swap would make sense (if it were just drive failure and not contents-scrambling.) I've worked on ~similar "tiny clean things" of one-off kind. Am aware of the nit-pick qualities of clean rooms, though I haven't been in a modern Super-nitpick variety. Just seems to me, that the risk/reward of an ad hoc kluge in a plastic bag: should suffice, for the dexterous manipulator of the common screws I see in drives I've disembowelled: there could be pretty good odds of success, enhanced by a final blowing out with super-clean Nitrogen, just as the cover goes back on. Quite harder if heads had also to be replaced; I grant that absence of suitable fixtures for that task, absent also the the lore of all those fuck-ups while learning -- does bad things to the odds. It would have to survive for a very short time, etc. What's to lose, especially if NONE of the data could be gotten off as-is. And you haven't a couple grand to give the white-coated ones at $500/hr inflated overhead. (This is all moot in the present example, of course - he says that DD is the operative tool == the sucker does its spinny/seeky thing.)
Post #299,224 by malraux 11/9/08 12:00:12 AM Reply	The data is scrambled. The local backup drive can find some things, not others. The one from chutney has a bad superblock or something. Regards, -scott <i>Welcome to Rivendell, Mr. Anderson.</i>
Post #299,227 by Ashton 11/9/08 12:40:39 AM Reply	Ah.. that's It then. It's all about TIME + expertise and patience. And everyone has a Life, which has no slot reserved for maddening fiddles with no withdrawal date. Fickle Finger Award (Laugh InÂ) to makers of that alleged array/controller. RAID, random assortment of indefensible drives? Lacking the requisite skillz, I shall at least loft a faint hope that one of our intrepid boffins can brilliantly deduce the exact madness-algorithm via which our orderly data came to look like the storage in Bush's brain. Sorry that our little fiefdom has now become {sob} an albatross. So, should we SHUN Mike? That seems to evoke magical powers.
Post #299,228 by folkert 11/9/08 1:40:35 AM Reply	Re: The data is scrambled. Can't use one of the backup Superblocks? There is usually at least 10... even on XFS.
Post #299,250 by malraux 11/9/08 10:30:12 AM Reply	The XFS utilities can't find them. Regards, -scott <i>Welcome to Rivendell, Mr. Anderson.</i>
Post #299,257 by crazy 11/9/08 11:37:59 AM Reply	I lost chunks of an XFS partition before I used to really like XFS for speed. Nothing could beat it in streaming, and random access was pretty good too. I ended up putting all my Oracle tablespaces on it, and then start using it for random other stuff. And then I got multiple corrupt partitions. I'm sure this was hardware. But the recovery (or failed attempt in some cases) of XFS partitions seemed far more painful than ext3. I ended up moving away from it. As I get older, it seems I prefer comfort over speed. In many areas of my life.
Post #299,258 by folkert 11/9/08 12:53:10 PM Reply	Bummer.
Post #299,267 by boxley 11/9/08 1:50:38 PM Reply	not familiar with the card, was it a hardware raid? Yes, Virginia, there really is an Obama Claus. You can hear him shouting, "On Pelosi, on Reid, on Kennedy, on Biden. Cash away, cash away, cash away all!"
Post #299,275 by malraux 11/9/08 2:15:36 PM Reply	Yes. Pain in the ass, too. The drivers caused kernel faults, and both times we lost a drive the mirror got trashed. Regards, -scott <i>Welcome to Rivendell, Mr. Anderson.</i>
Post #299,279 by boxley 11/9/08 3:17:33 PM Reply	that may be why the xfs tools cant find a superblock what was the name of the card again? I can do some poking around Yes, Virginia, there really is an Obama Claus. You can hear him shouting, "On Pelosi, on Reid, on Kennedy, on Biden. Cash away, cash away, cash away all!"
Post #299,280 by malraux 11/9/08 3:32:23 PM Reply	Re: that may be why the xfs tools cant find a superblock Believe it's a 3ware 9550, or something like that. I need to boot the machine to find out, but I'm trying to do that as infrequently as possible. Regards, -scott <i>Welcome to Rivendell, Mr. Anderson.</i>
Post #299,281 by boxley 11/9/08 3:34:23 PM Reply	thats fine, I can start there. (out of curiousity) is adaptec out of fashion? I suppose greg or andrew could answer that for me. Lately I need to get re-familiarized with intel based systems thanx, bill Yes, Virginia, there really is an Obama Claus. You can hear him shouting, "On Pelosi, on Reid, on Kennedy, on Biden. Cash away, cash away, cash away all!"
Post #299,287 by folkert 11/9/08 5:09:35 PM Reply	I ordered and installed the 9550. The 9550 was the only card available at the time that worked with the motherboard, that didn't cost $5K and was SATA. (Yes other boards were available, but the PCI spec on the Motherboard forced the choice... as it was currently not supporting certain types of cards) The Card's firmware needs to be updated seriously. The kernel Scott WAS running was Sarge's, Right? (2.6.8) There have been some firmware updates that have addressed many issues since I bought the card. As far as using Adaptec cards... FEH. I have MORE PROBLEMS, in both slowness and data consistency with SCSI and SATA via Adaptec RAID cards than the 3Ware Cards. Adaptec Firmware updates for "Adaptec brand" cards only, if you have an "OEM" card you have to get them from your OEM which drops your version 6 months after you buy the damn thing. I've only had ANY REALLY GOOD LUCK lately (4-5 years) with MegaRAID series of stuff... and I am still skeptical as they used to Sucksoplenty back when Dual Athlon MP processors were the rage... what 7-8 years ago? The only server class stuff I REALLY still trust at all is the IBM ServeRAID stuff. Bucksoplenty for them.
Post #299,291 by boxley 11/9/08 5:19:09 PM Reply	funny inside my ibm 3550's is all adaptec, why I asked Yes, Virginia, there really is an Obama Claus. You can hear him shouting, "On Pelosi, on Reid, on Kennedy, on Biden. Cash away, cash away, cash away all!"

Welcome to IWETHEY!