Server intermittently not responding, auth?

[Alexander Woodby <awoodby tir com>]

Ok, this server is the mail and dns server for a small ISP, also handles
radius.

Occasionally (every couple of hours now) it stops responding to outside
connections.

I can still log into it from console of course, and usually from the SGI
machine but not other hosts on the same network, nor will it respond to
radius requests, sendmail, or pop3.  It DOES respond to a ping though?

a netstat gives a LOT of 
tcp       11      0 bsd1.kode.net:auth      smv06.globecomm.n:63646
CLOSE_WAIT  
tcp       10      0 bsd1.kode.net:auth      165.251.8.74:42855     
CLOSE_WAIT  
tcp       10      0 bsd1.kode.net:auth      mail1.bellsouth.n:53790
CLOSE_WAIT  
tcp       10      0 bsd1.kode.net:auth      earth.voyageronli:42447
CLOSE_WAIT 


close_wait's mostly all pointing at auth processes.  It seems that my
pidentd may be messing up severely right?

Any ideas/tips/help?  

Thanks immensely!

--Alex