friedl@mtndew.Tustin.CA.US (Steve Friedl) (08/19/90)
Hi folks, I have been having the damndest time getting a customer machine back up after a crash this weekend, and I am baffled as to why it is not cooperating with me. They have a 3B2/600 running SVR3.2, and they have a 60MB cartridge tape and a pair of internal hard drives (all of this is SCSI). The root drive has a corrupted VTOC (probably my own damn fault), so I figured I would just boot from the OS tape and fix it right up. Fat chance. When I try to boot the machine, it does the SELF TEST and then hangs, presumably trying to filledt or run diagnostics off the hard disk. If I hit the reset, I get the traditional SYSTEM FAILURE HAS OCCURRED, but the only boot device it allows me is FD5 -- no SCSI choice. This is not a case of simply a missing name for an existing slot that would happen if filledt failed: the choices just aren't there. When I try to boot the essential utilities floppy to run filledt, it fills a few slots and then powers down the machine: great! I can't boot dgmon or /unix either. None of this changes if I reset the NVRAM first. What is happening here? I understand that maybe it can't boot from the hard disk, but why in the world won't it let me boot from tape? I really gotta get this machine working when I go out there on Monday morning at 6AM :-( Any help would be really appreciated. Steve -- Stephen J. Friedl, KA8CMY / Software Consultant / Tustin, CA / 3B2-kind-of-guy +1 714 544 6561 / friedl@mtndew.Tustin.CA.US / {uunet,attmail}!mtndew!friedl Combat global warming -- leave the refrigerator door open
kevin@cfctech.cfc.com (Kevin Darcy) (08/20/90)
In article <498@mtndew.Tustin.CA.US> friedl@mtndew.Tustin.CA.US (Steve Friedl) writes: >Hi folks, > > I have been having the damndest time getting a customer >machine back up after a crash this weekend, and I am baffled as >to why it is not cooperating with me. They have a 3B2/600 >running SVR3.2, and they have a 60MB cartridge tape and a pair of >internal hard drives (all of this is SCSI). > > The root drive has a corrupted VTOC (probably my own damn >fault), so I figured I would just boot from the OS tape and fix >it right up. Fat chance. > > When I try to boot the machine, it does the SELF TEST and >then hangs, presumably trying to filledt or run diagnostics off >the hard disk. If I hit the reset, I get the traditional SYSTEM >FAILURE HAS OCCURRED, but the only boot device it allows me is >FD5 -- no SCSI choice. This is not a case of simply a missing >name for an existing slot that would happen if filledt failed: >the choices just aren't there. > > When I try to boot the essential utilities floppy to run >filledt, it fills a few slots and then powers down the machine: >great! I can't boot dgmon or /unix either. None of this changes >if I reset the NVRAM first. > > What is happening here? I understand that maybe it can't >boot from the hard disk, but why in the world won't it let me >boot from tape? I really gotta get this machine working when I >go out there on Monday morning at 6AM :-( I'm not familiar with the "Essential Utilities Diskette" (I've always done such things from tape, or, when that fails, from the diskettes they distribute to data techs). I have, however, seen your power-off syndrome occur when techs try to run the wrong filledt/dgmon (the ones for smaller 3B2's). If you have a diskette that has utilities for ALL 3B2's on it, I think the ones for the 600 and above are usually prefixed with "600", e.g. 600filledt. I can't vouch for the names completely (it's been a while since I've had to perform this particular exercise - thank the gods). In any case, they have different names. You probably already know this, but you can get a list of files in the floppy directory by entering a non-existent file name at the "Enter path name" prompt. See what's there. The fact that you don't see ANY choices for SCSI is disturbing. I assume you've double-checked that the SCSI terminator is in firmly? If you were working on the inerds of the 3B2, you didn't happen to disconnect the cable that runs from the Host Adapter, by any chance? By the way, if you manage to filledt and still can't boot off the internal tape (assuming you get your SCSI choices back), try hooking up another tape drive externally and booting off that. I've found that marginal tape drives which work just fine for routine backup/restore stuff can nevertheless be lemons when it comes to booting the OS from tape. Don't ask me why. Some sort of weird timing consideration, perhaps? ------------------------------------------------------------------------------ kevin@cfctech.cfc.com | Kevin Darcy, Unix Systems Administrator ...sharkey!cfctech!kevin | Technical Services (CFC) Voice: (313) 948-4863 | Chrysler Corporation Fax: (313) 948-4975 | 27777 Franklin, Southfield, MI 48034 ------------------------------------------------------------------------------
tjt@cuuxb.ATT.COM (~XT6561340~Tom J. Toeller~C29~M19~6092~) (08/21/90)
In article <498@mtndew.Tustin.CA.US>, friedl@mtndew.Tustin.CA.US (Steve Friedl) writes: > When I try to boot the machine, it does the SELF TEST and > then hangs, presumably trying to filledt or run diagnostics off > the hard disk. If I hit the reset, I get the traditional SYSTEM > FAILURE HAS OCCURRED, but the only boot device it allows me is > FD5 -- no SCSI choice. This is not a case of simply a missing > name for an existing slot that would happen if filledt failed: > the choices just aren't there. > > When I try to boot the essential utilities floppy to run > filledt, it fills a few slots and then powers down the machine: > great! I can't boot dgmon or /unix either. None of this changes > if I reset the NVRAM first. > What you need is a giagnostic disk built on a similar machine, or the Maintenance Utilities floppy. This comes with the 3B2 Maintenance Reference Manual. The current release of the floppy is Rel 4.0. This floppy will work for any 3B2 (300-1000/80). There is no longer any reference to "600filledt" required. The boot block has the smarts to use the appropriate code for whatever hardware you have. Anyway, diagnostics on floppy will let you fiqure out why the SCSI bus is hosed. With this floppy, you can strip a version 3 machine down to just the system board and 1 memory board. About the power down problem! This will happen when you attempt to use certain versions of filledt from older machines. That is why, if you make your own dgn floppy, it is only good for like machines, or lesser machines of the same hardware release. For example, a floppy made on a 700 will work an another 700, a 600, 500, but not a 400 or 300. A floppy made on a 400 will work on another 400, a 300, but not a 500 or larger. The moral of the story is: Buy the book, get the floppy which works on all 3B2 machines, and you will have a great tool and a book for reference.
friedl@mtndew.Tustin.CA.US (Steve Friedl) (08/24/90)
In article <498@mtndew.Tustin.CA.US>, I write: > I have been having the damndest time getting a customer > machine back up after a crash this weekend, and I am baffled as > to why it is not cooperating with me. Hi folks, A few days ago I posted this request for help trying to find out why a 3B2/600 would not boot. I have fixed the problem and want to relay the things I learned. First, I hadn't known that the /filledt command on the Essential Utilities disks would not work on a 3B2/600 and will usually cause the machine to power off (bummer). The 3B2/400 is a version 2 machine, and their filledt programs are not equipped to handle the version 3 motherboards found on the /600. The filledt on the idtools disk (which I had at home) will work on all the machines. Second, the problem ended up being a bad tape drive (surprise!). After AT&T replaced the drive, however, we could no longer completely read either of our last two backups. We were in the middle of a repartition operation when the machine went down, so this was A Very Big Deal. The biggest bummer is that the bad spot on the main backups was in a 22MB file that was at the very end of the first tape. I took the tapes to my home and office machines and was never able to get everything off the tapes. I found that jiggling the tape in the drive would sometimes get past a bad spot here and there, but I could never get everything. What saved me was the AT&T Realtime Cartridge Tape Adjustment and Alignment Tool. To the untrained eye, this looks *exactly* like a pocket RS-232 screwdriver, and in fact you can convert a screwdriver to an adjustment tool just by calling it one :-) It turns out that wedging the screwdriver -- er, alignment tool -- between the top of the cartridge tape and the bottom of the upper end of the slot "misaligned" the tape enough that we were able to read all four tapes without a single bad spot. My customer suffered through two days of machine downtime, but we didn't lose any data and I ended up the tired hero. I want to thank the many people who wrote in with information, including the swell guy from AT&T (Doug Barnes?) who called me at 5:30 Monday morning before I went down to my customer. I hope I'm not missing anybody. Iain Bason M. Kosmin, LSI Communications Dave Bodenstab Doug Barnes AT&T DSG Janice L. Schlueter Kevin Darcy, Chrysler Financial Tom Toeller, AT&T DSG My customer and I thank you all. Steve -- Stephen J. Friedl, KA8CMY / Software Consultant / Tustin, CA / 3B2-kind-of-guy +1 714 544 6561 / friedl@mtndew.Tustin.CA.US / {uunet,attmail}!mtndew!friedl Q - Why do environmentals like fusion power so much? A - We don't have it yet