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

Re: Linux filesystems was [Re: Debian cd supporting ext4.]

Volkan YAZICI put forth on 7/27/2010 8:22 AM:

> You are missing a very important point: Durability to power failures.
> (Excuse me, but a majority of GNU/Linux users are not switched to a UPS
> or something.) And that's where XFS totally fails[1][2]. 

> [1] http://linux.derkeiler.com/Mailing-Lists/Debian/2008-11/msg00097.html

1.  Never quote forum or email posts as empirical or reliable evidence of
anything.  They are opinion, unless they quote fact from reliable sources to
back that opinion (which I did in my original post).  There are of course
exceptions to this rule of thumb, the main one being when the post is made by
a developer who is a recognized authority on the piece of software being
discussed.  In the case of XFS this would be Dave Chinner, Alex Elder, Eric
Sandeen, Christoph Hellwig, and others.  In the case of EXT2/3/4 this would be
Ted Tso, who happens, BTW, to work hand in hand with the XFS developers
because they rely on each others patches to other parts of the Linux kernel.
Not the last two entries:

In the case of Postfix this would be Wietse Venema and Viktor Duchovni.  In
the case of Linux this would be Linus Torvalds, Marcelo Tosatti, Alan Cox,
Andrew Morton and others.  Etc.

2.  If you are going to quote opinions from unreliable sources, at least read
the entire thread before quoting it.  In this case, again, your source
contradicts what you state, and then, oddly, himself:

"PS. Apparently they've improved some power-failure problems with XFS
since I used it, but then they also said that *before* I started using
it so I can't say I'd put any trust in that."

Aneurin Price admits he is aware of new changes that fix the problem, but then
discounts the fact and gives an anecdotal story as to why the fix isn't really
a fix.

> [2]

You didn't even read the information you linked to, which contradicts what you
state:  The fix for this issue has been in mainline since May 2007, over 3
years ago.

"Q: Why do I see binary NULLS in some files after recovery when I unplugged
the power?

Update: This issue has been addressed with a CVS fix on the 29th March 2007
and merged into mainline on 8th May 2007 for 2.6.22-rc1."

> And considering
> my personal experiences, reiserfs is the fastest fs (among ext3 and xfs)
> in terms of boot recovery phase times.

So, given that XFS recovery of any size filesystem is less than one second,
reiserfs recovery is so much less than 1 second that you notice the
difference?  From:

"XFS also provides file system journaling. This means that XFS uses database
recovery techniques to recover a consistent file system state after a system
crash. Using journaling, XFS is able to accomplish this recovery in under a
second, regardless of the file system size."

> [1] http://linux.derkeiler.com/Mailing-Lists/Debian/2008-11/msg00097.html

Here you quoted misinformation, because the opinions were based on experience
the OPs had with the software before it was patched to fix the problem, i.e.
more than 3 years ago.  You treated this as empirical evidence in your
argument against XFS.  I have shown this to be incorrect.

> [2] http://xfs.org/index.php/XFS_FAQ#Q:_Why_do_I_see_binary_NULLS_in_some_files_after_recovery_when_I_unplugged_the_power.3F

You quoted this FAQ item solely based on the tile, without reading it, in your
effort to denounce XFS.  The article clearly states the problem was fixed over
3 years ago, which you conveniently ignored.

>From now on, please get your facts straight, with proper documentation, before
trying to denounce a fantastic piece of FOSS into which many top-of-their-game
kernel engineers have put tens of thousands of man hours, striving to make it
the best it can be--and are wildly succeeding.

Join the xfs mailing list and you might learn something useful in place of
this trash you're talking about it.  Better yet, read what Hans Reiser had to
say about XFS.  He was totally enamored with it, and wanted to duplicate many
of its features:


Hans Reiser:
"XFS is an excellent file system, and there is an important area where XFS is
higher performance than we are...ReiserFS does a complete tree traversal for
every 4k block it writes, and then it inserts one pointer at a time into the
tree, which means that every 4k write incurs the overhead of a balancing of
the tree (which means it moves data around). For this reason, XFS has better
very large file performance...If you want to stream multi-media data for
Hollywood style applications, or use ACLs now rather than wait for Reiser4,
you might want to use XFS."

"This is an area we are still experimenting with. We currently do what ext2
does, and preallocate blocks. What XFS does is much better, they allocate
blocknrs to blocks at the time they are flushed to disk, and this allows a
much more efficient and optimal allocation to occur. We knew we couldn't do it
the XFS way...but reiser4 is being built around delayed allocation, and I'd
like to thank the XFS developers for taking the time to personally explain to
me why delayed allocation is the way to go."


Reply to: