[583] in linux-scsi channel archive
Re: 9GB drives
daemon@ATHENA.MIT.EDU (Drew Eckhardt)
Sat Sep 2 16:57:55 1995
To: linux-vger@wab-tis.rabobank.nl
cc: jduersto@kendall.mdcc.edu, willett@coam.usm.edu,
linux-scsi@vger.rutgers.edu
In-reply-to: Your message of "Sat, 02 Sep 1995 09:53:55 +0200."
<JAA25400@sys3.pe1chl.ampr.org>
Date: Sat, 02 Sep 1995 10:39:06 -0600
From: Drew Eckhardt <drew@poohsticks.org>
In message <JAA25400@sys3.pe1chl.ampr.org>, linux@pe1chl.ampr.org writes:
>According to Drew Eckhardt:
>> >If you're going to use the entire drive for linux, I don't see why you
>> >can't just "mke2fs /dev/sdb" and use the whole thing without a partition
>> >table. I posted a message about it before and nobody came up with any
>> >reason why this couldn't/shouldn't be done.
>>
>> You couldn't boot from it, would have one Big Fat Honkin partition
>> instead of multiple slices which are more desireable (to prevent
>> fragmentation on the main filesystme for instance, log files from
>> overflowing, etc), etc.
>
>Why can't you boot from it?
Because it would have blocks on it that were inaccessable to
the BIOS, which is used by LILO to load kernels.
>And isn't ext2fs supposed to prevent the fragmentation itself?
BSD FFS like filesystems are less likely to fragment than
traditional local unix filesystems which stick all of the files'
data blocks in the same place, but can still fragment; particularly
when you make lots of little files, delete, and make more big
files.
Putting the directories where this happens (like /tmp) on a different
partition avoids this.
