MMDF eating up cpu cycles, maxing out the machine

[DixieFlatline]

The problem stated a few days ago, when the deliver daemon wakes up to do a sweep, the load on the machine hits the roof.  Both sys and usr are maxed out yet vmstat shows normal swapping.  At first I though the queue is busy, I proceeded to do a checkque.  I discovered that checkque also maxes out the cpu.  I ran cleanque and cleanlck anyway.  They took forever to complete, slowing the machine down to a halt.  I notice no improvement in performance afterward.  I proceeded to kill and restart the deliver daemon.  Once again, it did nothing.  Does anyone have any idea what would cause this anomaly?

 

[DixieFlatline]

I found that while I was away, 2600 local msg has built up.  It is so big that the mail program is unable to clear the msg.  Is there a way to go into the q.local and erase the msg manually?  How do you tell the mail system to update itself if you do manually clear the queue?  Or is there a command that I can use the force mmdf to clear that specific queue?

 

[Guest_imported]

Same prob here. Anybody know anything about it?