Path: utzoo!utgpu!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!pasteur!agate!bionet!csd4.milw.wisc.edu!mailrus!cornell!uw-beaver!rice!sun-spots-request
From: bet@bent.mc.duke.edu (Bennett Todd)
Newsgroups: comp.sys.sun
Subject: Re: What is the best/safest way to fix a bad ND partition
Message-ID: <8901172013.AA05852@bent.mc.duke.edu>
Date: 23 Jan 89 22:48:05 GMT
Sender: usenet@rice.edu
Organization: Eikonix Corp., Bedford, MA
Lines: 15
Approved: Sun-Spots@rice.edu
Original-Date: Tue, 17 Jan 89 15:13:12 EST
X-Sun-Spots-Digest: Volume 7, Issue 116, message 5 of 13

The definitions for the partitions in /etc/nd.local can specify (in the
last column of the "user" lines) a "local ND" partition number for an ND
partition. Assuming you have assigned such numbers to your nd partitions,
you should be able to fsck the hosed partitions, and if that doesn't work
use one as a source and the other as a destination for a dd(1) command to
copy the partition. For the dd(1) approach I'd make sure I specified a
length (count=NN) for the copy operation; it may be that the nd(4) driver
correctly implements end-of-file -- but I don't know for sure, and would
tend not to trust it (how often would that particular feature be tested?).
Anyway, for a partition specified with "N" in the last column of the
"user" command in /etc/nd.local, that partition can be referred to as
/dev/ndlN (block device, for mounting and suchlike) or as /dev/rndlN (raw
character device, for fsck(8) and dd(1)).

-Bennett