OT - Journaling File Systems?

[Jesse Keating]

On Tuesday 27 April 2004 13:39, Edwards, Scott (MED, Kelly IT Resouces) 
wrote:
> My logic was just to use the closest thing to what we will be using,
> I.E. FC2.  I didn't think the FS stuff would be that affected.  I
> really didn't think it through.  My Bad.

I fell into that trap as well, and was doing performance testing on 
kernel 2.6 before I realized that things were in debug mode.  Test3 is 
a more appropriate platform, as it's more of an RC1 than a test 
release.

> >2) Fedora Core 2 Anaconda DOES support xfs/jfs/reiserfs.  You have
> > to pass these as options at boot time to the installer.  (linux xfs
> > jfs reiserfs)
>
> That works great!  I wish I had known that.  Is that documented
> somewhere that I missed?

I seem to remember them in the release notes, but I could be wrong.  It 
might be one of those "undocumented" features.

> >3) Fedora Core systems are build around ext3.  Usage of other file
> >systems is not tested very well, and unexpected problems may occur.
> >   This is not due to a file system problem, more of an operating
> > system not designed to use said file system.  To properly test, you
> > should test an operating system that targets said file system for
> > full operability.
>
> Ignoring #1 above for a moment, since we are using FC2 with whatever
> file system, it seems like it would be a waste to test them on some
> other Distro.  ;-)

Depends on what your goal is.  If your goal is to compare file systems, 
then you should use the appropriate OS.  If your goal is to compare 
Fedora Core 2 with different file system options, then yes, use FC2.  
But don't confuse FC2 results with general file system results.

> >4) NFS may not be an optimal test.  One would start with local disk
> >access, then extend it to other scenarios which may bring their own
> >instabilities into the picture.
>
> That's true.  It's ironic however that two of them didn't blow up
> until I started testing the local disk access.  Again my logic was
> just to test it in a manner as close as possible to how it will
> actually be used in normal operation.

That is a great target, once you have a baseline to compare against.  
Limiting the outside factors is necessary in order to generate the 
baseline, against which all other tests will be compared against.

Cheers!

-- 
Jesse Keating RHCE      (geek.j2solutions.net)
Fedora Legacy Team      (www.fedoralegacy.org)
GPG Public Key          (geek.j2solutions.net/jkeating.j2solutions.pub)
 
Was I helpful?  Let others know:
 http://svcs.affero.net/rm.php?r=jkeating
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: signature
URL: <http://listman.redhat.com/archives/fedora-test-list/attachments/20040427/d9ddac79/attachment.sig>