CentOS - Sep 2008 - CentOS 4 & autofs 4.1.3

For those of you using CentOS 4, is anyone having trouble with autofs 4.1.3? 
I'm seeing (with 60second unmount timeout) what appears to be an apparent
race condition where a script touches an automount directory at about (the same
second according to logs) the same time the unmount is triggered, and the script
then gets a read error.

Sure, increasing the unmount timeout will reduce the occurrence rate, but if
this really is a race, then it won't eliminate it.

In looking at the CHANGELOG between autofs 4.1.3 and 5.0.1rc2, there have been
tons of changes including some references to timing issues.  However nothing is
jumping out, and there are many changes.

Is it conventional wisdom that the 4.1.3 autofs has 'issues' and folks
routinely move up to autofs 5 on CentOS 4.x?

Thanks,

-John

John Kordash wrote:
>
> For those of you using CentOS 4, is anyone having trouble with autofs
4.1.3?
>  I'm seeing (with 60second unmount timeout) what appears to be an
apparent
> race condition where a script touches an automount directory at about (the
> same second according to logs) the same time the unmount is triggered, and
> the script then gets a read error.
I have it running on a couple hundred systems with no problems

nate