doc@holin.ATT.COM (David Mundhenk) (07/12/89)
Please don't flame me for cross-posting this! I need help and don't want to miss those who don't have unix-pc.general! -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- I suppose disk crashes are like door knobs - everyone gets a turn. Yesterday was mine. I hadn't had my system up in a few days, so I got a nice cold drink, sat down and hit the power switch. I wish I hadn't. (I know what happens when you ass-u-me, but I am led to believe this problem is disk related.) Just for info, I have a 7300 with 512K RAM on the motherboard and a 512K expansion for 1MB total. I have a 30MB CDC full height (~45ms.) drive of unknown history (it was used when I got it). The drive is powered by an external PC power supply (150W) since last winter my 7300 power supply decided it was overloaded. I think the last time I did a full restore was around last Dec. I have motherboard P3...P5 and am running UNIX3.5.1.4. When the machine started up, it put several rectangular boxes on the screen at the upper left, as always. Then little dots appeared all over the screen. After numerous tries, some of which yielded panic messages which I copied down but don't have in front of me right now, it finally booted. I promptly backed up some stuff I added recently and hadn't had a chance to yet. (whew!). Then something happened and /usr/lib/iv got trashed and I couldn't format more floppies. Things looked pretty messy, and I'm not fluent in fsdb, so I decided to do a full restore. (actually, on the next several reboot attempts, the system locked up after the message "real memory: .... available memory ....P3..P5"). The diags showed a read error at block 8104 (or so - don't remember) so I put that in the bad block list and reformatted. Restored the foundation set, dev set and a few other *bare necessities*. System came up just fine. Then I tried to restore my cpio backup (saved with cpio -ocvumB>/dev/rfp021 on 10 sector floppies). On disk #3, I got the message: "I/O failure on header: I/O error Can't read input; aborting." I tried to 'dd' in this floppy to poke around, and 'dd' said the same thing. W H A T G I V E S ? This is the first time I've had trouble with a cpio backup on this system. Could it be a bad floppy, are my heads dirty, or what? Is there a way to fix this? I know which files are in this section of the backup, and they aren't critical, so if there is a way to bypass this section somehow, I'm all ears! (I tried cpio with a "!pattern" at the end - it still bombs). Should I be making TWO backups of important stuff? I'll sh*t a golden brick if I can't restore these files! Actually, most of them were sources from USENET and friends, so with some blood, sweat, tears, and bended knees I can probably replace them... P L E A S E H E L P ! Any and all replies will be received with great joy and rapture! Thanks for all your help - everyone has been tremendously helpful before. I am a very regular reader of this newsgroup, and I know some of you gurus will probably have easy solutions to my plight. (Lenny, Thad, John, & others, are you listening?) Thanks! Dave Mundhenk :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: EMAIL: ...!att!holin!doc | "I can't complain but | /^, VOICE: (201)-580-4943 | sometimes I still do"| / } _, , , __ #include <std.disclaimer> | - Joe Walsh | /_./ (_l |/ <~_ :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: -- :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: EMAIL: ...!att!holin!doc | "I can't complain but | /^, VOICE: (201)-949-5308 | sometimes I still do"| / } _, , , __ #include <std.disclaimer> | - Joe Walsh | /_./ (_l |/ <~_
cals@cals01.NEWPORT.RI.US (Charles A. Sefranek) (07/19/89)
In article <588@holin.ATT.COM> doc@holin.ATT.COM (David Mundhenk) writes: > > ... Then I tried to restore >my cpio backup (saved with cpio -ocvumB>/dev/rfp021 >on 10 sector floppies). On disk #3, I got the message: > >"I/O failure on header: I/O error >Can't read input; aborting." > ... > >W H A T G I V E S ? > >This is the first time I've had trouble with a cpio backup on this >system. Could it be a bad floppy, are my heads dirty, or what? > I ran into this problem too! It drove me nuts until I finally found out what was going on. I remember other people on the net complaining about this too, usually blaming cpio. I should have suspected something when I bought a BIG batch of those cheap floppies and hardly ever found a bad one... If you format floppies with the ua (and even if you don't) BE CAREFUL - the "surface test" it runs DOESN'T ALWAYS FIND BAD FLOPPIES!!! Even using iv won't find them. The only reliable method I've found is to use iv with the -l option to force 10 passes of the surface test. This takes a LONG time but its worth it. The command to use is: iv -iwl /dev/rfp020 descriptionfile (See the manual for descriptionfile). Naturally you have to follow this with the appropriate mkfs command if you want to put a file system on it (not necessary for cpio diskettes). After I discovered this, I wrote my own shell script to format floppies and guess what -- I found LOTS of previously "good" floppies that are actually bad. I went through all my backups weeding them out. Keep an eye on the /usr/adm/unix.log file when you are formatting floppies. If you format a floppy that has never been formatted before you should typically get four entries for it. If you reformat a previously good floppy, you should get NO entries. If you get entries for a previously formatted floppy, and it appears to format OK, then it encountered soft errors during the formatting - I would put it aside and only use it for non-critical info. The system will automatically delete the unix.log file if it gets too big (somewhere around 10K bytes). Oh, by the way, if you understand how to create the descriptionfile per the manual above, you can format floppies with 10 sectors per track, and 42 cylinders; that's right 42 cylinders, not just 40!!! This gives me an extra 40 blocks on each floppy. Can't guarantee it'll work on everybody's drives, but I know at least two other people who do the same. (Don't try for 43 cylinders, the drive makes an awful crunch when it hits the stops!) -- Charlie Sefranek cals@cals01.NEWPORT.RI.US UUCP: {rayssd,xanth,lazlo,mirror}!galaxia!cals01!cals Alt.: c4s@rayssdb.ray.com {sun,decuac,gatech,necntc,ukma}!rayssd!rayssdb!c4s