[Beowulf] OS for 64 bit AMD

Joe Landman landman at scalableinformatics.com
Fri Apr 1 22:44:25 PST 2005

Hi David:

   I wasn't challenging you, I just had trouble understanding some of 
what you were saying.

David Kewley wrote:
> Thanks for your challenges, Joe.  Let me give you a bit of my background.
> I am at this moment building a NAS server, and am working to enable and test 
> xfs on RHEL4, for a ~8TB data volume.  I *want* to try xfs on RHEL4, and I 
> hope to use it in production if I can satisfy myself that it's the right 
> choice.

Use FC-3++ and install with the xfs option.  This will allow you to 
format your large area as either ext3 or xfs.

> I get challenged by my colleages to justify straying from the default 
> filesystem (ext3) that RH supports.  I hope to give them good reasons, 
> because so far I like the feel of xfs (I have used it in a couple of other 
> projects).  The final choice is mine, but I feel a responsibility to answer 
> my colleagues' challenges.
> The thread has gotten too tangled, so I'm going to drop most of it and make 
> some short statements etc. instead.
> You say xfsdump works, as if regular dump doesn't.  Dump *does* work.  What is 
> so much better about xfsdump?

Actually I was being a bit flippant when you asked what was meant by 
xfsdump working.  Guy Coates mentioned that ext3 had some issues.

I have had problems in the recent past (early february) with a drive 
that required some dumps.  This disk had both xfs and ext3 file systems. 
  I was able to get reads of some of the partitions.  I could not 
recover the ext3 file system no matter what I tried (dump and quite a 
few other things).  I was able to recover the xfs file system fairly easily.

YMMV.  Mine did.

> I'd really like to see comparisons of TODAY's xfs with TODAY's ext3, using 
> whatever tests or comparisons you find most meaningful. 

Last I checked, my systems are running todays versions of everything.  I 
don't necessarily consider steaming bits in CVS/subversion to be todays 

> I'm not very 
> interested in benchmarks or application tests from the past.  I am making a 
> choice between xfs and ext3 today, not last year, and given that ext3 is 
> claimed to have improved very recently, I want to see if those improvements 
> change the balance.

Sounds like you need to take real programs, do real runs, and make real 
timing measurements.

> Given my current work, I may have some (artificial) benchmarks in a week or 
> so.  I would very much welcome suggestions about what to try, especially 
> given yours (and other list members') expertise in this area.

I would suggest that you get the applications list from your users.  As 
many have stated in the past, and many will again state in the future, 
the only meaningful benchmarks are your real codes.  Anything else 
generates waste heat in the universe.

> RHEL4 beta versions were available for public download starting 9/27/04; 
> interested parties could have started making detailed comparisons then.
> Even though I wish xfs was supported in RHEL4, I'm not convinced that RH's 
> choice *not* to support xfs is ill-considered or short-sightedly 
> self-interested or done only out of competitive considerations.

If you read the forums as you pointed out later, you will note that 
there has been a large chorus for quite some time about this.  Redhat's 
response to this chorus has been "wait, ext* will get better".  Yes, it 
has gotten better, but no, this is not what the customers asked for.

Whether or not it is ill-considered is a personal opinion.  I believe it 
to be.  Self-interested it most definitely is (remember what it invests 
in).  All companies must, by definition if they plan to exist for any 
length of time, be self-interested.

 > I find the
> argument reasonable that RH chose to stick with ext3 because they believe it 
> meets all their customers' needs, 

I am astounded... when the customers in the threads you mention, and in 
the preceding threads made it clear over the last few years that it did 
not indeed meet their needs, and can they please have support for 
something that does ...  I don't think that is a reasonable argument 
that what they provide is good enough and meets needs, when clearly 
specific large customers are asking for something different.

> and they don't want to invest in talent to 
> support other enterprise filesystems for major support-purchasing customers.

Again, if small one-man distros can support the system, the argument 
rings hollow.

> In particular, see this thread and its continuation in the October archives:
> https://www.redhat.com/archives/nahant-beta-list/2004-September/msg00049.html
> Even if you don't read anything else from this thread, check out these two 
> posts from Arjan van de Ven, which quite forcefully presents his position on 
> ext3 versus xfs:
> https://www.redhat.com/archives/nahant-beta-list/2004-September/msg00058.html
> https://www.redhat.com/archives/nahant-beta-list/2004-September/msg00057.html
> If you believe he's misguided, argue with what he wrote. :)

Don't need to.  This has been debated in various threads for the last 4+ 
years.  Ext3 will continue to improve, and it is probably good enough 
for some section of users.  For others, who need what xfs provides, ext3 
is not sufficient.


>>>Yup.  They don't have to support it, which is likely why they don't
>>>bother disabling it. :)
>>Oddly enough, it works quite well in FC.  Imagine that.  They can
>>have their cake and eat it too.  Without spending time/money/effort. 
>>Other distros are doing it, and they are too in FC-x.
>>Only they choose not to in RHEL4.  Obviously support has nothing to
>>do with it (FC-x is a beta program for RHEL, and you need to fix your
>>customers bug reports to make it a useful program).
> How can you say, "Obviously support has nothing to do with it"?  I don't get 
> that.  They have major customers buying expensive support for RHEL.  They 
> offer no paid support for FC.  If they choose, they can ignore bug reports on 
> FC xfs.


Redhat pushes them off to the fs maintainers, as they should.  Bugs 
ignored make for bad practice.

If it were a support issue for RH, it should be a support issue for 
SuSE, for Mandrake, for Debian, for ....

But it isn't.

> David

Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web  : http://www.scalableinformatics.com
phone: +1 734 786 8423
fax  : +1 734 786 8452
cell : +1 734 612 4615

More information about the Beowulf mailing list