Managers:
I had asked
> But when I tried to mount the root partition, the OS panic'ed with:
>
> > Aug 11 20:19:22 zantar vmunix: ADVFS EXCEPTION
> > Aug 11 20:19:22 zantar vmunix: Module = ftx_recovery.c, Line = 198
> > Aug 11 20:19:22 zantar vmunix: ftx_bfmeta_rec_redo: got bmt page N1 instead
> I relabeled the "bad" disk and installed Tru64 Unix successfully. So it seems
> that there was nothing wrong with the disk hardware, just data corruption.
>
> My question: what could be on a disk so that Tru64 panics instead of just
> failing to mount the disk?
As Dr. Tom Blinn <tpb_at_doctor.zk3.dec.com> says
> the answer is simple: a corrupted AdvFS file system.
or in somewhat more detail alan_at_nabeth.cxo.dec.com explained
> To answer the question; very wrong data. I have to suppose
> that as part of being a log-based file system that doesn't
> require fsck, AdvFS has to assume that between the data
> in the log and on the file system that it is consistent.
> If it couldn't make that assumption, then it would have
> to check the consistency at each mount, which would be
> much like having to run fsck; so much for that marketing
> advantage.
>
> On the other hand... It has to replay the log as it mounts
> the file so it can get back to what it hopes is a consistent
> file system. In the course of this reply it probably makes
> some consistency checks. If the call path is unique for
> mounting, then it might be quite reasonable to refuse to
> mount the file system instead of panic. If the path is
> shared with that of normal I/O, then we have to ask whether
> this particular corruption is worthy of a domain panic or
> a whole system panic.
Scott Balfour <scott.balfour-eds_at_eds.com> had similar bad experiences with AdvFS
panics.
I had also mentioned
> As an aside, is anybody else as upset as I that after paying $30,000 for a
> computer including $500 for a mandatory FIS which needs to be redone anyway,
> Compaq cannot include a $5 operating system CD?
Ian Mortimer <ian_at_physics.uq.edu.au> seems to speak for many when he replied
> Yes. Lot of people it would seem judging from comments I've
> seen on this mailing list and others.
--Harvey
----
Harvey Rarback phone: (650)926-3963
Stanford Linear Accelerator Ctr fax: (650)926-4100
PO Box 4349, MS 69 home phone: (650)560-9111
Stanford, CA 94309-0210 Web: ssrl.slac.stanford.edu/~rarback
rarback_at_ssrl.slac.stanford.edu ESnet: 45746::RARBACK
"Always tell the truth, that way you don't have to keep track."
Pogo
Received on Mon Aug 16 1999 - 23:17:43 NZST