PEPRBV@CFAAMP.BITNET (Bob Babcock) (06/26/87)
I just installed a replacement hard disk in a Mizar system running OS-9/68K. The manufacturer's tests show one bad sector, but this is not detected by the verify stage of format, so the sector is not locked out. Mizar's documentation says to use the Mizar badsector utility in a case like this, but it turns out that they haven't yet written this utility, and can only recommend that I run format repeatedly in hopes that it will eventually find the bad spot. I could easily see wasting more than a man-day learning enough about the file structure and writing custom code to solve this problem, which means that it might be cheaper to buy another disk. Has anyone solved this problem already? Reasonable solutions would be to figure out the mapping between head/cylinder number and OS-9 sector number and fiddle with the bit map, or thoroughly test all the unallocated disk sectors and leave a read-only file on top of any bad sectors found. Thanks for any suggestions.
berger@datacube.UUCP (06/26/87)
This should be considered a bug. There should be a way to enter the
manufactures defect list into the format utility. There should also
be a place on the disk where it could be recorded in a machine readable
form.
Bob Berger
Datacube Inc. Systems / Software Group 4 Dearborn Rd. Peabody, Ma 01960
VOICE: 617-535-6644; FAX: (617) 535-5643; TWX: (710) 347-0125
UUCP: berger@datacube.COM, ihnp4!datacube!berger
{seismo,cbosgd,cuae2,mit-eddie}!mirror!datacube!bergerpete@wlbr.UUCP (Pete Lyall) (06/29/87)
In article <PEPRBV.870625.215954.A0@CFAAMP.BITNET> PEPRBV@CFAAMP.BITNET (Bob Babcock) writes: >I just installed a replacement hard disk in a Mizar system >running OS-9/68K. The manufacturer's tests show one bad sector, >but this is not detected by the verify stage of format, so the >sector is not locked out. Mizar's documentation says to use the >Mizar badsector utility in a case like this, but it turns out >that they haven't yet written this utility, and can only >recommend that I run format repeatedly in hopes that it will >eventually find the bad spot. I could easily see wasting more >than a man-day learning enough about the file structure and >writing custom code to solve this problem, which means that it >might be cheaper to buy another disk. Has anyone solved this >problem already? Reasonable solutions would be to figure out the >mapping between head/cylinder number and OS-9 sector number and >fiddle with the bit map, or thoroughly test all the unallocated >disk sectors and leave a read-only file on top of any bad sectors >found. > >Thanks for any suggestions. Well, we have a program written by an OS9 UG member (actually, a staffer as well - Carl Kreider) that wrote something called HDREAD. It reads a disk a track at a time, and then backs up and isolates bad sectors if a track has a hit. You can use two companion utilities with this: a) Hooz - determines who the sector's owner is (if any), and b) Mapout - which toggles the sector bit in the bitmap. These may be set up in filter/pipe fashion (i.e. hdread -h ! mapout -z). Caveats: a) I'm not sure that this code is truly PD - I will check. b) It was written & used under OS9/6809 level II, but since it's in MW C, it should be trivial to port. Carl has both 6809 and 68000 systems. -- Pete Lyall Usenet: {trwrb, scgvaxd, ihnp4, voder, vortex}!wlbr!pete Compuserve: 76703,4230 (OS9 Forum Sysop) OS9 (home): (805)-985-0632 (24hr./1200 baud) Phone: (818)-706-5693 (work 9-5 PST) EATON Corp, IMSD, 31717 La Tienda Dr., Westlake Village, Ca. 91359 ----------------------------------------------------------------------