home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!munnari.oz.au!mips!mips!odin!fido!zola!zuni!anchor!olson
- From: olson@anchor.esd.sgi.com (Dave Olson)
- Subject: Re: updating bad block list on SCSI disk using fx
- Message-ID: <p5lvogo@zuni.esd.sgi.com>
- Sender: news@zuni.esd.sgi.com (Net News)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <1992Aug28.191401.29136@unixg.ubc.ca>
- Date: Sat, 29 Aug 92 03:05:20 GMT
- Lines: 52
-
- In <1992Aug28.191401.29136@unixg.ubc.ca> laplante@ocgy.ubc.ca (Denis Laplante) writes:
-
-
- | I tried to use the fx command 'auto' to reformat a disk suspected of having
- | bad blocks. The man page says
- | auto Initializes a new disk. The disk is formatted, a label is created
- | and written to it, and it is exercised to detect and map out bad
- | blocks.
- | Fx then printed 213 cryptic 2-line errors, all the same except for a block
- | number near the block in error, in the form:
- | fx/auto/complete: warning - data mismatch
- | fx/auto/complete: Error 0 -- Couldn't find bad block on track starting with \
- | block 283590 after 2 tries
-
- That means you either have bad cabling/termination problems, or a drive
- with extremely flaky firmware. fx won't forward the bad blocks unless
- it is convinced that they are really bad; it isn't able to find the
- failing blocks when it goes to look for them. You could use -r0 to
- suppress some of the retries, which might find some of them.
-
- | 1- How likely are those 213 ignored bad blocks to come back and haunt me?
- | Will these blocks just invisibly suffer from automatically recovered
- | errors?
-
- They aren't really 'bad' in the sense of causing an error, but they
- aren't returning the same data that was written to them, at least
- some of the time, which is definitely going to cause problems!
-
- | 2- Is there a way to log recovered SCSI disk errors, as a warning of trouble
-
- Yes; use the /label/set/param menu to enable reporting of recovered
- errors; the kernel will print them, and syslogd will log them to
- /usr/adm/SYSLOG
-
- | 3- What did I do wrong for the bad block list to not be updated?
-
- There weren't any hard errors.
-
- | 4- Is there a way to declare new bad blocks without losing data, and without
- | I would guess that on an unmounted partition the following might do it:
- | badblock/addbb
- | debug/writebuf (the same block address now accesses a new disk sector)
-
- As of 4.0.5, fx will attempt to read the data before forwarding the
- bad blocks on SCSI drives (it always has for other drive types), and
- if successful, will rewrite the data after the block is forwarded.
-
- I'm not sure that will fix your problems though.
- --
- Let no one tell me that silence gives consent, | Dave Olson
- because whoever is silent dissents. | Silicon Graphics, Inc.
- Maria Isabel Barreno | olson@sgi.com
-