xfs
[Top] [All Lists]

RE: Data Corruption Problem

To: mahesh.babbar@xxxxxx
Subject: RE: Data Corruption Problem
From: Net Llama! <netllama@xxxxxxxxxxxxx>
Date: Wed, 30 Jul 2003 12:35:06 -0400 (EDT)
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <H00000bb0040b8ce.1059579394.dlx101.dlh.st.com@MHS>
References: <H00000bb0040b8ce.1059579394.dlx101.dlh.st.com@MHS>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Wed, 30 Jul 2003 mahesh.babbar@xxxxxx wrote:
>      Hi Lonnie
>
>      I am sorry If I was being offending, but want to present following
>      facts here. Please challenge me :
>
>      Evidence is
>
>      1. I have lost a big chunk of data. In the range of GBs. It was
>      commercially important data.

That's what backups are for.  Your lack of backups isn't anyone's fault
but your own.  Your lack of maintaining a system with a recent filesystem
codebase is no one's fault but your own.

>      2. In my 10 years of carreer, I have never seen a Sun(UFS) or an
>      HPUX(JFS) or an AIX box(JFS) corrupting data like what I have seen or
>      what Aman has faced.

So?  You also didn't think it was prudent to keep backups of your data.
So you're background isn't exactly something that I'd want to brag about.

>      2. You will say "LATEST" code resolves most of the issues. Accepted.
>      But is there a version, X or Y or Z (old or latest, a user doesn't
>      care) which is STABLE.

That's what the 2.4.x kernels are.  They are the stable kernels.

>      What I mean by stability is, a technical person (which you all are)
>      can say " I am confident that IT WORKS".

I've got terrabytes of data on XFS.  None of it has ever failed or been
lost.  But if it did, i have nightly backups, and could restore withou a
problem.  And since, as someone else noted, XFS is 100% free, i've got no
place spreading FUD, or complaints.  If XFS isn't meeting your needs,
either make a positive effort to fix any documented, genuine problems that
exist, or don't use XFS.  You are free to throw more money away on any of
the commerical Unixes, if that's what makes you sleep better at night.

>      3. In an envrionment which involves *real* customers and where each
>      second counts and data means hard dollors, no body bothers about CODES
>      and their versions. So please don't distribute something (even free)
>      in the name of
>
>      * it's latest
>      * It's gives better performance
>      * It's this , it's that
>
>      BUT
>
>      when it comes to stability, It all depends. Which as per me is the
>      most important factor.
>
>      It may work on our personal laptops/labs/academies but not in
>      commercial sector.

Please do not condescend to me, or anyone else about real customers.  95%
of the data that I host on XFS is customer & production data.  I'm talking
nearly 3TB of data.  If you a sysadmin, its your job to maintain the
server, and ensure that its running the latest stable codebase so that
uptime & stability are maintained.  Making up excuses isn't going to
change anything.  And latching onto someone else's report of a problem is
really nothing more than bandwagon hopping.

Either provide documentation of a real stability problem with XFS in
2.4.21 kernels, or stop spreading FUD.

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Lonni J Friedman                                netllama@xxxxxxxxxxxxx
Linux Step-by-step & TyGeMo                  http://netllama.ipfox.com


<Prev in Thread] Current Thread [Next in Thread>