------- Comment From <email address hidden> 2015-11-03 20:56 EDT-------
Thanks for chasing down the missing dependency side of the equation here.
That still leaves the potential of deadlock if the callout exec ever returns non-zero for any other reason. Are the risks considered low enough to simply document this as a potential hazard of writing process accounting information across an NFSv4 mount?
------- Comment From <email address hidden> 2015-11-03 20:56 EDT-------
Thanks for chasing down the missing dependency side of the equation here.
That still leaves the potential of deadlock if the callout exec ever returns non-zero for any other reason. Are the risks considered low enough to simply document this as a potential hazard of writing process accounting information across an NFSv4 mount?