Alessandro Selli wrote:
>
> Which is a use that goes beyond what an LPIC-101 Minimally Qualified
> Candidate is supposed to know.


Not making an argument either way, however ...

A filesystem mounted over another filesystem is an extremely common (and
poor) situation in any POSIX environmemt.

One thing that continues to bother me in these discussions are the fact
that the debates are over commands and inclusion preferences, instead if
concepts and the tools that could be used to identify and/or solve them.

Again ... not weighing in on the tool, but a filesystem mounted over a
filesystem is an extremely common, poor, elusive and troubling scenario for
1st line Linux sysadmins.  I've seen it handled very poorly by 1st level
sysadmins too.

Matt:  Is there are "reference history" on why debugfs was included?  That
might go a long way to explaining it's inclusion, and the justification for
Level 1 as well.

I recently read Red Hat dropped support for BTRFS:
>
> https://servers-linux.ro/red-hat-to-drop-support-for-btrfs-linux-magazine/
> Red Hat to Drop Support for Btrfs ยป Linux Magazine
> 09/08/2017


Clarification:  Red Hat _never_ supported btrfs.

Unlike most Linux distributors, Red Hat is extremely anal on the "support"
term in it's "entitled" products. That's because whatever Red Hat
"supports," it supports for 10+ years in many cases.

Insert a lot of messy scenarios for various distributors that left a lot of
customers in poor situations -- sans SuSE who really expends a lot to make
good on its life cycle promises as well.

File systems are one of those things where various distributors ship and
support all sorts of options, and Red Hat does not.  Btrfs had been in
perpetual "tech preview" and Red Hat finally "yanked the plug" on most of
it's engineering efforts --finally advising customers it hadn't reached the
level maturity amd stability desired, and it would not for the foreseeable
future.

It still ships in Fedora, but the "tech preview" status is over in Red Hat
Enterprise Linux (RHEL), having never gone to release status. This is rare,
but has happened before. When it goes to "release," Red Hat is "married to
it." In btrfs' case, it never made it to the wedding.

Are there major distributions that do not use ext4 as the default
> filesystem at installation time on servers, desktops and laptops?


Yes, several have installers that still select Ext by default.

And x86-64 (aka AMD64, also IA-32e) is often required for XFS too, so x86
(aka IA-32) doesn't include it. This has to do more with the userspace
tools.

Ext is not going anywhere anytime soon.

That, plus, Ext (especially Ext4) and XFS are still extremely similar to
each other in concepts. Other than xfs_repair (and xfs_rsr), most of the
tools work the same.

 I agree here too


I meant to bring this up earlier, but it's more important to understand
what each tool does, is used for and addresses than debating the tool
itself.

I.e., tools that let sysadmins view super block and inode information is
the focus.

Sysadmins who dont understand the concepts and what is contained in the
super block and inodes are going to have a much rougher time -- even at
Level 1 -- than ones who do.

The classic ... "Why is 'df' [always] so fast and why is 'du' usually much
slower (at least the first time)?"

That, plus, "What is this 'superblock' thing that the error message pointed
out"?

As I said, I've very worried people are debating commands and preferences,
instead of the concepts addressed with the tools.

Maybe Matt can provide some history and/or other references to explain why
various tools were included?

- bjs



-- 

-- 
Bryan J Smith  -  http://www.linkedin.com/in/bjsmith
E-mail:  b.j.smith at ieee.org  or  me at bjsmith.me
_______________________________________________
lpi-examdev mailing list
[email protected]
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev

Reply via email to