kpropd deadlocking?

[Jack Neely]

Greetings,

I'm using krb5-server-1.6.1-36.el5_5.5 from RHEL 5.  Of 3 kerberos
slaves, one of them seems to continuously have the kpropd daemon
deadlock.  I'm out of ideas and looking for help.

A script on the master runs kprop and pushes to all my slaves.  The
slave in question will eventually (within a few hours) accept a kprop
request, fork off a child, receive data, build the from_master.temp
file, and hang part way through.  The from_master.temp isn't complete,
the kpropd child process is stuck on a read() syscall.  The kprop client
on the master is stuck on a write() syscall.

The kprop client on the master eventually times out, but on the kerberos
slave the kpropd child waits forever.  Of course, future kprop attempts
fail because one is already in progress.  The master pushes runs the
kprop script every 5 minutes.

I've replaced the hardware.  The slave is identical to the one sitting
right beside it (physically and network wise) that is not having issues.  

Could this have something to do with Nagios monitoring the kprop port to
frequently?  Any help would be appreciated.

Jack

-- 
Jack Neely <jjneely at ncsu.edu>
Linux Czar, OIT Campus Linux Services
Office of Information Technology, NC State University
GPG Fingerprint: 1917 5AC1 E828 9337 7AA4  EA6B 213B 765F 3B6A 5B89