[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Possibly head-crashed disk

  Date: Tue, 16 Aug 88 16:59 EDT
  From: Guy Caruthers <guyc@PEGASUS.SCRC.Symbolics.COM>
  Subject: ECC error (actually, disk space recovery)

    Date: 12 Aug 88 18:38 EST
    From: STERRITT%SDEVAX.decnet@ge-crd.arpa

[...my whole description deleted...]
    incremental world.  Oh yeah, it's a T306 drive, if that helps.
	    chris sterritt

    Date: Fri, 12 Aug 88 19:35 EDT
    From: Scott McKay <SWM@SAPSUCKER.SCRC.Symbolics.COM>
    Subject: ECC error (actually, disk space recovery)

[...Scott suggests that I read the documetation on si:fix-fep-file...
BUT IT ISN'T DOCUMENTED in any of the hardcopy, at least that I could find!
Perhaps he meant the document examiner?  We don't have 7.2's docs up yet.]

    Date: Sun, 14 Aug 88 21:39 EDT
    From: barmar@Think.COM
    To: STERRITT%SDEVAX.decnet@ge-crd.arpa

[...barmar reminds us that worlds with bad blocks aren't worth much...]

Finally (so far) Guy states:

  A T-306 disk pack should be guaranteed media (ie: 100% error free).
  If you are repeatedly experiencing a problem in a specific area, then
  the pack is no longer error free.  The solutions mentioned above may
  provide short term relief, but I would recommend having your CSE
  come out and try to determine why the area is bad.  It could be 
  something simple like a header block that was written bad.  It
  could also have been caused by a soft head crash.  If it is the 
  latter, the sooner you find out the better, as your problems
  will spread to other areas.  If there is no indication of a
  csrah, then formatting the pack should recover the bad block,
  (not something you like to hear about).  

Well, before I got this I tried si:fix-fep-file.  It said that there was
a read error on that block, and asked if I'd like it to try a write-read
test.  I said yes, and it did it, and evidently that fixed the problem;
it said that there was now no problem, and trying si:fix-fep-file again
found no read error.  So I deleted the world holding that block and rebuilt
the world.  All went well, and I can gc again.

NOW, the question is: I guess the right thing to do is watch carefully for
further errors, and if I get them, call my CSE.  Is it serious enough (we've
had no other errors with that drive) to call him immediately?  Reformatting
the pack is really no problem, as it only holds paging files and worlds.