From: ruediger@ramz.ing.tu-bs.de (Ruediger Helsch) Subject: Re: Pt 1/6: SLS 99.2 problems Date: 2 Feb 1993 00:39:44 GMT
In article <1993Jan28.183547.27995@mark-one.inet-uk.co.uk>, mark@mark-one.inet-uk.co.uk (Mark Powell) writes:
|> I had the same problem just today. First time I've seen a Kernel Panic from
|> linux. "Kernel panic: unable to read i-node block". Reset the system and
|> booted from a1/a2. Ran efsck once and it spewed out lots of the usual stuff.
|> Ran it a second time and no problems (just a check). Rebooted from the hard disk
|> and the same panic appeared after the kernel had started up (after all the SCSI
|> stuff, but just before, I presume, init had started up). Went back to
|> the disks and tried to mount the partition. Fine. Had a look around and all the
|> files appeared to be there. "Great", I thought just a matter of a back-up to
|> tape and mkefs and re-install. However, the tape just refused to go. Must
|> be something wrong with the kernel on A1 as I use tapes routinely from the
|> 0.99pl2 kernel.
The same thing happened to me. I found that something with the inodes
was really bad, as the system crashed as soon as I dumped (or listed) some
of the recently active files. Fortunately it was possible to make a tar with
the v option, noting the files causing the crash and avoiding them after the
next floppy boot. A tedious procedure, but saved me half a weeks work.
If it matters, it happened with the newest (0.99.4) kernel version.
Btw, I am surprised that efsck seems to do something in your case. As often
as I used it, it never seemed to do anything than display long lists of these
<0 / >MAX_BLOCKS (or so) error messages.
Ruediger Helsch <ruediger@ramz.ing.tu-bs.de>