Subject: Re: Questions about 1.3
To: None <current-users@NetBSD.ORG>
From: der Mouse <mouse@Rodents.Montreal.QC.CA>
Date: 10/02/1997 22:35:38
>> 1. Will 1.3 still have the limit of 8 partitions / disk (on i386
>> port specifically)?
> Personally, I don't find 8 partitions limiting at all.
Neither do I...on some systems. On others, it borders on crippling.
> IMHO, subdividing a disk into more than one partition per file system
> is a bad way to go,
I don't think _anyone_ has suggested using more partitions than
filesystems. (I've done it, it's true, but only under relatively
special circumstances; I can't recall seeing it ever suggested here as
a good thing to do, never mind seeing it so suggested in this thread.)
> because then you run into problems with some partitions filling up
> quickly and running out of room while other partitions still have
> plenty of space, and you have to work around it by creating symlinks
When this happens, it is not always a "problem" that needs "work[ing]
around"; indeed, sometimes it is exactly the desired behavior, and can
be the very reason the disk was partitioned as it was: to prevent one
thing eating space from running some other thing out of space. Not all
machines - not even all NetBSD machines - are home boxen with only one
or two users and no question of one thing running another unexpectedly
out of space.
> The ccd concept, I think, was a Very Good Idea.
It was, and is - for some environments.
More than 7 partitions per pack is another Very Good Idea - for some
There may be some overlap between these two "some environments"s,
though probably not much - but regardless of how much or little overlap
there is, there are still good reasons to do both.
> Provided the cluster/fragment sizes remain optimal, we need fewer
> partitions per drive, not more.
Who's this "we"? It certainly doesn't include me. I've already been
glad I have support for >8 partitions per pack on the sparc and sun3
7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B