trumpour@uvicctr.UUCP (trumpour) (04/21/87)
My attempts to boot 4.3bsd off a "direct" (ie NOT mapped as an RM05) Fujuitsu 2351A or 2361 Eagle (SI 9751 or 9761) attached to a VAX 11/780 via a Systems Industries 9900 controller fails with the message: "unknown drive type". I assume that this is because the programs in /sys/stand, as provided by Berkeley do not know about the SI drives. SI did provide an updated vaxmba/hp.c with appropriate header files and several useful utilities for formatting and repairing their disks--all running under UNIX. Unfortunately, they do not appear to have provided the software necessary to boot UNIX up in the first place! Has anyone else run into this problem and found a solution (other than configuring the SI 9761 as two RM05's)? Allan Trumpour ...{uw-beaver,ubc-vision}!uvicctr!trumpour
mangler@cit-vax.Caltech.Edu (System Mangler) (04/25/87)
Berkeley doesn't have any SI controllers, so they haven't maintained the parts of hp.c that relate to SI controllers. They haven't added a SI9761 case to the switch statement, so you won't be able to boot that drive, at all. The case for SI9751D (direct mapped, RM05 drive type) is still there, and you can try booting that. But 4.3's error recovery code does not work at all on SI controllers. If you have a bad block in the swap area, like we did, you won't get very far. We were lucky; we had a friend who could compile a kernel for us. You should set jumper W7A on the motherboard for Internal ECC, and find an Eagle whose first 84,000 sectors can all be read correctly with just hardware ECC. The remaining errors will have to be covered up by badsect(8) until you get the driver compiled, like in the days before BAD144. If you get a bad block in an inode area, try telling newfs to allocate fewer inodes, or change the cylinder group size. Don Speck speck@vlsi.caltech.edu {seismo,rutgers,ames}!cit-vax!speck