[conspire] verifying I'm doing this correctly - fdisk, mkfs.ext3...

[Rick Moen]

Quoting Don Marti (dmarti at zgp.org):

> And some people are adamant about turning on noatime
> everywhere.

True.  But...

> Or am I missing something?

I don't know what to say, except to suggest that it's a good idea to
disclose "requires atime updating on [dirspec]" in the docs; you could 
also check filesystem support at startup time and advise the user if
atime isn't being updated.

Traditionally, it's all been a free-for-all, but sysadmins know they
screw with filesystem options at their peril, and might break something.
The fact that distros do _not_ disable atime/diratime means the burden's
basically on them -- in theory, anyway.  But I'd probably put in a few
lines to check atime functionality after about bug report #2.  ;->