Re: file systems
On Sat, Apr 30, 2011 at 11:43 PM, Stan Hoeppner <email@example.com> wrote:
> On 4/26/2011 5:58 PM, shawn wilson wrote:
>> you know, i don't mind religious debates - vi vs emacs, mac vs
>> windows, iphone vs android, fibre vs iscsi, trustedbsd vs selinux,
>> and... xfs vs ext. however, i like an educated debate where both sides
>> can site specifics and hard facts and not just say 'this is better
>> because i say so' (yeah, stan, you got a bit more specific, but not
> This thread wasn't actually a debate about filesystems. It was about an OP
> ditching a quality FS for a truly silly reason.
'quality FS' - opinion.
'silly reason' - opinion.
> If you're interested in learning about the merits of each XFS, search the
> list archives, google, and hist the XFS website. There's plenty of
> information available. I've listed many/most of the benefits of XFS on this
> list many times.
i'm interested in not seeing unsubstantiated opinion on a technical
to be fair, you've listed a few unbiased references (i'd consider
xlf's site somewhat biased). first, p psu paper on high bandwidth fs's
(damn good reading), a kernel.org doc on xfs (dry, skimmed it), and
wikipedia on ext4 (didn't read it).
> Default filesystem choice by a distro team has much more to do with
> familiarity, continuity, and politics than features, performance, or
> limitations of an FS. Note that Novell made the decision to switch SLE[SD]
> from Reiser to EXT3 very shortly after Reiser killed his wife.
we can have a debate on why sles is about the shittiest distro on
earth if you'd like. i would be more than happy to go through the
files under /etc they ship with and point out all of the useless,
disgusting crap they put in there if you'd like (at least as of 2
years ago). in other words, i don't think i'd use them as a reference
for anything good or proper.
> Debian team members have quipped in the past that Debian will ship with EXTx
> as the default until the end of time, currently determined to be 2038. ;)
you might be right about epoch. but we'll be bogging down the 4to6
gateways long before then. maybe we should fire icann and ietf (or is
it ansi or itu) for being stupid? :)