named consuming 99%+ CPU on both NS1 ans NS2 simultaneously

[naidyhorski at home.com]

I am experiencing a problem that I dont have an answer for. I am
seeing CPU spikes reaching 99%+ on both the primary and secondary
nameservers simultaneously. The servers spike at the same time and
return to normal (1% to 7%) at the same time. If I delete all zone
files on NS2 and restart it, it will create copies of all zone files
without any problems. If I kill named on NS2, NS1 runs like a champ
with no CPU spikes. The interval between spikes is irregular.

I was originally running DNS services on Solaris 2.6...then 7. We
recently replaced the Sun Sparcs with 2 Intel based servers running
Red Hat 6. I simply transfered the named.conf and db.* files to the
new nameservers and brought them online. All seemed to be well...until
I noticed slow response times when browsing....dug alittle deeper and
found that named was consuming all my resources :( These particular
servers are authorative for about 20 domains.

The old servers were Sparc based with 500Mhz processors w/ 512 megs of
memory...ran like a champ!! The New servers are Dell Poweredges with
PIII 500s and 256 megs of memory. They shipped with Red Hat.

Anyone heard of any reasons why named would cause such spikes?

Stephen J naidyhorski
naidyhorski at home.com