On Tue, 4 May 2004, Jeff Moyer wrote:

> ==> Regarding [autofs]  mount point created if automount map is empty during daemon 
> startup?; Tom Georgoulias <[EMAIL PROTECTED]> adds:
> 
> [snip]
> 
> tom.georgoulias> * Thu Mar 18 2004 Jeff Moyer <[EMAIL PROTECTED]> -
> tom.georgoulias> 1:4.1.1-3
> 
> tom.georgoulias> - Fix bug in get_best_mount, whereby if there is only one
> tom.georgoulias> option, we choose nothing.  This is primarily due to the
> tom.georgoulias> fact that we pass 0 in to the get_best_mount function for
> tom.georgoulias> the long timeout parameter.  So, we timeout trying to
> tom.georgoulias> contact our first and only server, and never retry.
> 
> Ugh, skeleton is out of the closet.  ;-P  To clarify, I only saw this when
> the map was an LDAP map (which I thought should make absolutely no
> difference).  This patch is actually just a work-around.  I was hoping to
> track down the real issue later...

Saw that patch. Was a bit curious about it. Didn't add it.

I think that I discovered that the error return from get_best_mount was 
not being checked. This may have been the cause of that.

I thought that was fixed.

CHANGELOG says 4.1.2.

I think the replicated server behaviour is working OK now. Fingers 
crossed.

Ian

_______________________________________________
autofs mailing list
[EMAIL PROTECTED]
http://linux.kernel.org/mailman/listinfo/autofs

Reply via email to