Subject: bin/11112: newfs_lfs should deal with -m 0 differently
To: None <gnats-bugs@gnats.netbsd.org>
From: None <gendalia@iastate.edu>
List: netbsd-bugs
Date: 10/01/2000 07:01:25
>Number:         11112
>Category:       bin
>Synopsis:       newfs_lfs -m 0 raid0d results in equivalent of newfs_lfs -m 10
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    bin-bug-people
>State:          open
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Sun Oct 01 07:07:00 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     Tracy J. Di Marco White
>Release:        1.5 branch, cvs update Sept 30, 2000
>Organization:
>Environment:
        
System: NetBSD lyra 1.5_ALPHA2 NetBSD 1.5_ALPHA2 (LYRA) #1: Sat Sep 30 17:52:36
CDT 2000 root@lyra:/mnt/src/sys/arch/i386/compile/LYRA i386

>Description:
newfs_lfs -m 0 raid0d results in 10% reserved space.  tunefs suggests that
setting to 0 should be possible.  (although tunefs doesn't work on lfs
filesystems)

In general, using integer percentages for reserved space wastes a lot of
space on larger disks, but this is a problem across everything that uses
-m #.
>How-To-Repeat:
newfs_lfs -m 0 disk
mount disk
df
/dev/raid0d    83765470      901 75388022     0%    /mnt2

>Fix:
workaround:
newfs_lfs -m 1 raid0d results in the least reserved space:
/dev/raid0d    84757227      723 83908932     0%    /mnt2
>Release-Note:
>Audit-Trail:
>Unformatted: