[OpenAFS-win32-devel] URGENT: 1.3.72 to release this Sunday

[Kris Van Hees]

Answers inlined below...

On Thu, Oct 14, 2004 at 01:40:13PM -0400, Rodney M Dyer wrote:
> At 11:41 AM 10/14/04, Kris Van Hees wrote:
> 
> >As far as I know, the problem that is being triggered on the Unix server 
> >side is SMP-dependent, and has most likely been present in all OpenAFS 
> >releases up to (and including) 1.3.71.
> 
> Let me get all my eggs in one basket here.
> 
> 1.  It's the server daemon that is crashing?

Correct.

> 2.  The OpenAFS Windows client 1.3.71 and previous versions with crypt 
> turned on will tickle the server into crashing?

Correct.

> (Jeff, I already understand that the Windows client can crash if run on a 
> hyperthreaded system with 1.3.71 and earlier.  I'm only concerned about the 
> servers.)
> 
> 3.  People who are running Unix/Linux etc, don't run 1.3.71 (this is a 
> Windows version), they are running 1.2.11 right?

Usually, yes.  Some brave souls run 1.3.71 (or even CVS HEAD), either out of
some necessity (2.6 kernels) or just for testing.  Production use is still
usually 1.2.11.

> 4.  As far as servers are concerned, is this a PC SMP only issue in Linux 
> or does it also occur on Solaris SUN machines?

Any SMP box.  I know of this problem happening both on Linux and Solaris.

> Quite frankly, our AFS servers have never crashed while running 1.2.10/11 
> in the last year (barring a bad disk), and on the clients we are using a 
> mix of hyperthreaded and non-hyperthreaded PC machines without crypt turned
> on.
> 
> Rodney
> 
> 
> 
> _______________________________________________
> OpenAFS-Win32-devel mailing list
> OpenAFS-Win32-devel@openafs.org
> http://lists.openafs.org/mailman/listinfo/openafs-win32-devel

-- 
Never underestimate a Mage with:
 - the Intelligence to cast Magic Missile,
 - the Constitution to survive the first hit, and
 - the Dexterity to run fast enough to avoid being hit a second time.