dbr@cybvax0.UUCP (Douglas Robinson) (11/29/84)
I have just become aware of a very nasty little bit of incompatibility. If you have heard of this before, please ignore this - we are relatively new to the net. We just finished formatting a new Fujitsu Eagle / SC750 combo and learned the hard way (MANY wasted hours) that: 4.2BSD wants the BAD SECTOR information in the MANUFACTURER'S AREA, NOT in the SITE AREA where the DEC and EMULEX formatters put them!!!!!!!!!! 4.2BSD does NOT recognize the information put down by the DEC and EMULEX formatters! You have to print them out using the DEC/EMULEX formatter and then enter them by hand via bad144(8). This little tidbit of information appears NOWHERE in the documentation that was distributed to us with 4.2BSD!!! NOT only that, but when bad144(8) runs, it wipes out the information in the SITE AREA (zero's it). (We learned this due to a new formatter from EMULEX that prints out the two areas separately!!) -- Doug Robinson Jobs don't kill programmers... programmers kill jobs! Cybermation, Inc. 617/492-8810 377 Putnam Avenue Cambridge, MA 02139 (USA) ...!{mit-eddie, harvard, mirror}!cybvax0!dbr
root@uokmet.UUCP (12/04/84)
> 4.2BSD wants the BAD SECTOR information in the MANUFACTURER'S AREA, NOT > in the SITE AREA where the DEC and EMULEX formatters put them!!!!!!!!!! > 4.2BSD does NOT recognize the information put down by the DEC and EMULEX > formatters! You have to print them out using the DEC/EMULEX formatter > and then enter them by hand via bad144(8). This little tidbit of > information appears NOWHERE in the documentation that was distributed to > us with 4.2BSD!!! The same problem exists on 2.9BSD. From bad144(8): "The bad sector file information is located in the first 5 even numbered sectors of the last track of the disk pack." Our Emulex diagnostic pack puts the info on the 2nd 5 even numbered sectors. The solution was to modify "rm.c" (rm02 driver), badsect(8) and bad144(8). The modification for "rm.c" is below. Some drivers (including the "rm") may need the change in more than one place... bbp->b_bcount = sizeof(struct dkbad); bbp->b_un.b_addr = (caddr_t)&rmbad[unit]; bbp->b_blkno = (daddr_t)RM_NCYL * (RM_NSECT*RM_NTRAC) - RM_NSECT + 10; /* Emulex diagnostic starts the info on sector 10 */ The change is similar in bad144.c (in the "lseek" calls). Kevin W. Thomas Univ. of Oklahoma School of Meteorology Norman, OK 73019 UUCP: ...!ctvax!uokvax!uokmet!kwthomas (root)
scottha@azure.UUCP (Scott Hankerson) (12/05/84)
In article <248@cybvax0.UUCP> dbr@cybvax0.UUCP (Douglas Robinson) writes: > >4.2BSD wants the BAD SECTOR information in the MANUFACTURER'S AREA, NOT >in the SITE AREA where the DEC and EMULEX formatters put them!!!!!!!!!! >4.2BSD does NOT recognize the information put down by the DEC and EMULEX >formatters! You have to print them out using the DEC/EMULEX formatter >and then enter them by hand via bad144(8). This little tidbit of >information appears NOWHERE in the documentation that was distributed to >us with 4.2BSD!!! Instead of using the DEC or EMULEX formatter, use the 4.2 formatter. The 4.2 formatter forwards the bad blocks the correct area for 4.2BSD. I'm surprised that the DEC formatter doesn't put the bad blocks in the same place, especially when the place that Berkeley puts them corresponds to the DEC standard 144.
sid@linus.UUCP (Sid Stuart) (12/18/84)
>In article <248@cybvax0.UUCP> dbr@cybvax0.UUCP (Douglas Robinson) writes: > >4.2BSD wants the BAD SECTOR information in the MANUFACTURER'S AREA, NOT >in the SITE AREA where the DEC and EMULEX formatters put them!!!!!!!!!! >4.2BSD does NOT recognize the information put down by the DEC and EMULEX >formatters! You have to print them out using the DEC/EMULEX formatter >and then enter them by hand via bad144(8). This little tidbit of >information appears NOWHERE in the documentation that was distributed to >us with 4.2BSD!!! > All right guys, I am confused. I formatted my Eagle with the Emulex formatter when I was running 4.1. I switched over to 4.2 without reformatting my disks and they still work. Has emulex come out with a new gonzo formatter? Doug, are you sure of the procedure you used to format the disks? This situation sounds a little bizzare to me, does anyone have any more in depth information before I spend a weekend reformatting my eagle just to see what happens? Does anyone have the source to the formatter? Berkeley seems to have ommited it on our distribution. sid at linus doug, I live at 617-271-2839 I would be interested in finding out more about this.