This seems to be a common issue, but I’ve not seen a complete answer.
I have an Intel based MT device with two interfaces - one public, one private. DHCP is used on the inside network and the DNS server IP provided to clients is that of the internal MT interface, as follows:
[iain@fw-stj-02] ip dhcp-server> print
Flags: X - disabled, I - invalid
# NAME INTERFACE RELAY ADDRESS-POOL LEASE-TIME ADD-ARP
0 dhcp-stj08 inside pool-sjt08 8h yes
[iain@fw-stj-02] ip pool> print
# NAME RANGES
0 pool-sjt08 xxx.yyy.zzz.50-xxx.yyy.zzz.80
My DNS settings on the MT device are as follows:
[iain@fw-stj-02] ip dns> print
primary-dns: isp.dns.ip.1
secondary-dns: isp.dns.ip.2
allow-remote-requests: yes
cache-size: 2048 kB
cache-max-ttl: 7d
cache-used: 55 kB
What I am finding is that when using ‘dig’ a DNS request from an internal client can take 1-2 seconds to resolve. But If an applicaiton (web/email) tries to contact a server the rsolution takes ~10 seconds, even if there is a static entry in the DNS cache…which, BTW, only ever holds the static entries!
Does the DNS service work correctly with MT??
Regs.