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