NCP Exclude IP Address - Unreachable IP address causes slowness

Error -625
Network performance is slow
Server communication slow
Drive Mapping is slow
Log in is slow
DS Synchronization slow
Time Synchronization slow
Report Synchronization Status slow
Can't Ping IP address
NWAdmin object search is very slow
NCP Exclude IP Address - Unreachable IP address causes slowness

NCP Exclude IP Address

Trouble shooting steps:
1) Identify non-reachable areas of your network.  You should be able to find your unreachable IP address by using one or more of the following:
a. Review network topology documentation
b. Run DSRepair on the server you are connected to - Advanced Options | Servers known to this database (select the slow server) | Repair all network addresses.  This will return -625 errors for each address that is not reachable from the server that you ran the repair on.
c. DSTrace - This needs to be run on the server that you are connected to, which does not hold a copy of the object you are trying to resolve.
At the console type:
set dstrace = on
set dstrace = nodebug
set dstrace = +rn
set dstrace=*m9000000
set ttf=on
set dstrace=*r

Run tests that show the slowness then stop the trace file logging by typing:

set ttf=off

DS errors like -625's will identify servers that are not reachable.

d. LAN Trace - Taken from perspective of the server.  The intent is to capture NCP requests to other servers to identify which server cannot be reached.

2) Once you have identified an unreachable IP address, go to the server that owns the unreachable address.  You can then either use Monitor or a Set command to exclude this address from accepting NCP traffic.
a. Load Monitor | Server Parameters | NCP | NCP Exclude IP Address
b. Set NCP Exclude IP Addresses = <the unreachable IP address>

3) This change should be immediate but DS may need time to process the change.  If it doesn't seem to be taking effect (meaning this server still hands out the unreachable address) try initiating a limber process DSTRACE *L.  If that doesn't work you can try a reinitialize server which should signal to DS that a change has occurred with respect to bound IP addresses.  If neither of these help, a reboot will always make the NCP Exclude take effect.

Any of the symptoms above can be caused by an unreachable IP address on the network that is being advertised (over NCP via an NDS referral list) as a source for information. The address may be an isolated backup segment, a public address on a Border Manager server, etc.

 When a device tries to communicate with an unreachable IP address that has come back in a DS referral it will try until a time out occurs (approximately 1 min) and then it will try another IP address (from the referral list).

This attempt to establish a connection with the unreachable IP address until the timeout occurs causes the delay which is the root of the slowness.

The next address attempted, if reachable, will then provide the necessary information and the process will succeed.
Formerly known as TID# 10067926