On Tue, 2015-12-01 at 15:31 -0600, Chris Adams wrote:
Once upon a time, Patrick O'Callaghan
<pocallaghan(a)gmail.com> said:
> Because I know what physical disks I have in my machine and I want
> to
> relate that to what I see in the output of df. I might even want to
> move a device to another machine and be able to mount the right
> partitions in the right places. With "normal" (i.e. non-LVM)
> partitioning it's fairly easy to do this. With LVM it's definitely
> not.
I find quite the opposite: without LVM, I have to know that the drive
I just moved from computer to computer changed from sdb to sdc, and
edit fstab and such manually.
Maybe so, but it's still easy to do.
With LVM, I still get /dev/vg_foo/lv_bar, and
don't care what raw device the underlying partition is, how it is
connected, etc. (very useful for example when taking an internal
drive from one computer and connecting it via an external adapter of
some type on another).
Which is fine if a) the second machine also runs LVM (what if it's on
an Ubuntu machine without LVM, rather than Fedora?) and b) the two use
the same LVM logical layout.
I guess my point is that the average Linux admin is going to have a
working knowledge of disk partitioning, whereas LVM is an *additional*
layer of expertise that may pay dividends in certain use cases, but for
most people is just irrelevant.
Anyway, we're getting way off the original topic of this thread. I
didn't really want to start a whole discussion (all of which has been
said before more than once).
poc