[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: partman-zfs warning template wording



Quoting Robert Millan (rmh@debian.org):

> Hi Christian,
> 
> > If I understand well, partman-zfs is unstable on kfreebsd-i386 but
> > fine on kfreebsd-amd64. However, would it be a problem to talk about
> > "32 bits systems" and "64 bits systems" instead of "i386" and "amd64"?
> 
> This is a common missconception.  ZFS has stability and performance
> problems on Intel 386 due to limitations of this architecture, it's not
> because of word length. I386 carries on with 30 years of ill design
> decisions and legacy baggage. Other 32-bit architectures such as
> ARM or MIPS don't suffer from this kind of issues.

My main point is to avoid "jargon". Here, we are talking about
kfreebsd ports where only two flavours exist: -i386 and -amd64. Both
run on the same family of processors and the main difference between
them is that one is 32-bits and the other is 64-bits.

So, as what we basically want to say users is "hey, if you pick ZFS
file system, you guys should really pick the so-called amd64
architecture", my proposal seems to be a good way to do so.

Maybe, if we want to be picky, we could say:

" Although ZFS is supported on 32 bits i386 family systems, using it
without special tuning may lead to performance or stability problems."


> 
> Specifically, the problems have to do with small number of
> general-purpose registers and with inability to implement
> privilege separation at the MMU level.
> 
> > These arches names are confusing enough, imho.
> 
> What's confusing about them?


I think that a random poll among users would easily lead to many
people thinking that "amd64" should only be used with systems with AMD
processors. This is what I call "confusing".




Attachment: signature.asc
Description: Digital signature


Reply to: