Date: Mon, 07 May 2001 23:57:31 +0200 From: "Julian Stacey Jhs@jhs.muc.de" <jhs@jhs.muc.de> To: freebsd-scsi@FreeBSD.ORG Cc: pechter@ureach.com, "Ron Klinkien" <rklinkien@chello.nl> Subject: Re: Recovered data with positive head offset field replaceable unit msg Message-ID: <200105072157.f47LvVs42257@jhs.muc.de> In-Reply-To: Message from Bill Pechter <pechter@ureach.com> of "Sat, 05 May 2001 07:55:49 EDT." <200105051155.HAA22713@stage21.ureach.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Bill Pechter wrote: > > Hi SCSI wizards, > > > > On one of my FreeBSD machines I found > > this rather strange scsi message: > > > > > (da1:sym0:0:8:0): READ(06). CDB: 8 8 1 90 10 0 > > > (da1:sym0:0:8:0): RECOVERED ERROR info:80198 asc:17,2 > > (da1:sym0:0:8:0): Recovered data with positive head offset > field replaceable unit: df sks:80,12 > > > > Thanks in advance. > > > > Regards, > > Ron. > > Haven't seen something like that in years. > In the old days of SMD disks and DEC RP06's and RM05's (mid > 80's), I'd have said it was a servo tracking issue with the > heads missing the data by a bit (possible bad spot on disk) > which was recovered by the disk controller offsetting the heads > a bit towards the center of the drive just a bit. > > The data was recoverable... by the drive moving the head just a > little. Things like this were caused by bad spot on disk (servo > platter or data platter) misalignment, dirt on the drive rails > that caused the heads to bounce a bit while seeking or trying to > stay on track, power fluctuations. Some idiot banging into the > disks during a read operation. I'd heard (pre '85 I think) that if someone deleted data, & a country's spy service wanted to get the data back, they `just' recalibrated the seek offset servos, increased the op amp gain factor, & read as before, (or repeat scanned if necessary) to recover at least most of the deleted data. There are normal occasions when the owner too wants to recover his own data. Anyone know if SCSI (&/or IDE come to that) provides hooks [doc'd or undoc'd ] ? It's possible to recover most text data from bad floppies by repeat scanning, even without varying the offset, even with failing CRCs: I wrote a program that does that in ~87 http://bim.bsn.com/~jhs/src/bsd/jhs/bin/public/valid/valid.c & valid.1 (The recover part of the program's now larger functionality runs only on DOS though, not on Unix, as DOS passes filled buffers back even when read() returns CRC error, whereas FreeBSD discards the buffer content on error. Julian - Julian Stacey Unix Consultant - Munich Germany http://bim.bsn.com/~jhs/ Ihr Rauchen => mein allergischer Kopfschmerz ! Kau/Schnupftabak probieren ! Like Linux ? Then also look at FreeBSD with its 5000+ packages ! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200105072157.f47LvVs42257>