home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.apple2
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!swrinde!gatech!destroyer!cs.ubc.ca!alberta!glyn
- From: glyn@cs.UAlberta.CA (Reade Glyn Devin)
- Subject: help -- problems with hard drive
- Message-ID: <1992Dec30.201622.19839@cs.UAlberta.CA>
- Sender: news@cs.UAlberta.CA (News Administrator)
- Nntp-Posting-Host: manning.cs.ualberta.ca
- Organization: University of Alberta, Edmonton, Canada
- Date: Wed, 30 Dec 1992 20:16:22 GMT
- Lines: 80
-
- Hi folks.
-
- I've got a problem with my hard drive that I haven't been able to
- track down, and I would appreciate any suggestions. I'm using a 40 Meg
- AE Vulcan Gold upgrade, ZipGS 9/32, Prosel 16 with error recovery and
- disk verification on backup. There are (normally) two partitions, the
- first one being 25024.0k. Memory is a Ramkeeper with 3Meg GSRam+ and
- 500k apple card. Other cards are normally used, but were removed for
- trouble-shooting. Here's the story:
-
- The problem first came up while trying to do a complete-mode
- optimization of my first partition using ProSel.16. The error was
- a block reference error.
-
- I did a restoration from backup (I keep a series of three backups
- for each partition, used in succession). I encountered an error $27
- (I/O error) on the restoration.
-
- By the way, I always do a directory repair & bad block lockout (Prosel
- volume repair) before any restorations from backup.
-
- I tried again starting with a low level format by typing 'AE' at the
- main menu in Partition Manager (Vulcan Gold version). It reported the
- bad tracks of $15.0 $1A.2 $1B.2 $1C.2
-
- In addition to the (three) bad blocks reported by Prosel (2762, 2773,
- 5955), I also locked out the blocks mentioned in the paragraph above.
- I think I translated them correctly as:
- 2267, 2268
- 2888, 2889
- 2996, 2997
- 3104, 3105
- I included two blocks each to account for a possible off-by-one error.
-
- I no longer got error $27 on restoration, but every time I try a complete-
- mode optimization I get 'Block Reference Error, Reference Address $0CCD'
- I've tried this multiple times with different backup sets. I have tried
- locking out any index blocks used by the file (and its parent directories)
- being optimized at *crash* time. The same error (Reference Address $0CCD)
- always occurs even though different blocks are being optimized at the
- crash time. $0CCD is also locked out. There are no problems when I
- do an optimiztion in Turbo-mode (turbo-mode does not optimize directories
- or index blocks, just data blocks).
-
- I thought that maybe I had a bad set of initial blocks on my partition
- (blocks 0-5 and the bit-map), so I tried making a small '/BAD' partition
- with my normal partitions following. Only when /BAD is 1.5Meg or larger
- do the problems disappear. When the size of /BAD is changed, the blocks
- assigned to BAD.BLOCKS are recalulated with the offset. When resizing
- /BAD, I was careful not to get any known bad blocks in the first $12 blocks
- of the partition.
-
- An interesting thought is that regardless of the size of /BAD, the error
- is _always_ "Reference Address $0CCD". If $0CCD was a bad block, you
- would expect the reference to change with the size of /BAD. Since index
- files store addresses with the (reversed) high and low bits offset by
- $100, I don't have a utility which can tell me which blocks make
- reference to $0CCD (remember that since $0CCD is locked out, Prosel
- Info Desk will report it as part of BAD.BLOCKS).
-
- I lock out blocks out by using Prosel 16 Zap Block Editor (Block Warden),
- modifying the BAD.BLOCKS file (only file on disk at the time) then using
- directory repair to fix the volume bit-map.
-
- I've verified that all software in use (Prosel, Partition Manager) match,
- byte-for-byte, with the original disks.
-
- There is no possibility of virus-infected software being involved at
- this stage, unless it can hide itself in the BRAM (which I seriously
- doubt).
-
- I checked my ROM disk (now being used as a boot volume) and RAM using
- the Ramkeeper CDA. MemoryTest.CL reports no errors on RAM. There is
- no RAM disk.
-
- As you can see, I'm grasping at straws. Any help appreciated.
-
- --Devin glyn@cs.ualberta.ca
-
-
-