Preinciple server works; Clone duplicate does not. Why?

carlfoss at world.std.com carlfoss at world.std.com
Wed Oct 6 17:42:14 UTC 1999


I have been constructing a clone of my primary server that has been working
for more than a year.  As far as I can tell, WinNT/TCPIP/DNS are set up
identically on both.  Yet when I attach the principle server to my ISP it
works, and when I remove the principle and attach the clone instead, it
does not work.  It is close.  *Both servers will successful do pings and
traceroutes to outside servers.  However, the principle server will respond
to pings and traceroutes, while the clone won't respond to either.*  I am
trying to understand why this might be, and what to do to remedy the problem.

All the files such as named.boot, dn.inaddr, db.cache are identical.  I
copied them from one server to the other.  I am also running the same
version of DNS Bind on both computers.  The principle is using WinNT v 4
with SP3, while the clone is using WinNT v4 with SP5.  The principle is
using OReilly Web Site v 1.something, while the clone is using v 2.x.
Version = named 4.9.5-Rel+-Monday-11-November-96
	GregSchueman-LarryKahn-VirajBais

I managed to uncover the following discrepancy between named.run when I
start up (named -d 1) the respective DNS servers while connected to the
internet. (See last two lines, although the third lines are also very
interesting.)

for the server that does not respond to pings from outside:
rrsetupdate: I.ROOT-SERVERS.NET
rrsetcmp: name not in database
rrsetupdate: I.ROOT-SERVERS.NET 0
13 root servers
qfree: ns E.ROOT-SERVERS.NET rcnt 1 (busy)
qfree: nsdata 192.203.2

for the server that does respond to pings from outside:
rrsetupdate: C.ROOT-SERVERS.NET
rrsetcmp: name not in database
rrsetupdate: C.ROOT-SERVERS.NET 0
rrsetupdate: G.ROOT-SERVERS.NET
rrsetcmp: name not in database
rrsetupdate: G.ROOT-SERVERS.NET 0
13 root servers
qfree: ns E.ROOT-SERVERS.N

Earlier in the files there is a set of what appear to be non-substantial
discrepancies: for example
considering [109.11.89.146]
dqp->dq_addr 109.11.89.146 d_dfd 36
listening [109.11.89.146]
Whereas for the non-working clone, the 36 is replaced by a 60. (Note that I
have changed the IP address above from the real one for security reasons.)

Does anyone have an idea as to what is amiss here, and how I might go about
fixing it?


More information about the bind-users mailing list