Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!comp.vuw.ac.nz!actrix!paul
From: paul@actrix.gen.nz (Paul Gillingwater)
Newsgroups: comp.unix.sysv386
Subject: mkfs under 386/ix
Message-ID: <1990Dec4.080453.3739@actrix.gen.nz>
Date: 4 Dec 90 08:04:53 GMT
Organization: Actrix Information Exchange
Lines: 29

Yesterday I had to rebuild a file system on our 386/ix 2.02 system
for the second time.  The file system is mounted as /usr/spool/news,
and needed about 10,000 more inodes.  :-)

The problem:  the standard ISC docs donfor 2.02 't have a mkfs man page.  
The man pages I could find were different (from non-ISC source).
I couldn't work out how big the file system originally was, in
blocks, because I'd lost the mkfs.data that was originally created.

How should one find this out?  Specifically, how can I empirically
establish the total size available for an EXISTING section (I won't
call it a partition, because several sections may share the same
entry in the partition table which is accessed by fdisk).

My concern is that when i used 

mkfs /dev/dsk/0s3 117000:28000 1 2048

I got exactly the size in blocks that I requested.  My worry is that
I don't know how big it could be before it starts to overlap onto
the next disk section.

I chose 117000 because it was somewhat less than the original size
on blocks for the section.  28000 is the new number of inodes,
necessary for a healthy news feed here.  1 is the inter-sector gap I
think -- we're using an ESDI controller for this drive.  2048 -- I
wasn't sure whether this should be something else...
-- 
Paul Gillingwater, paul@actrix.gen.nz