Notes/Domino 4 and 5 Forum

trying to debug this cluster issue

I have two Domino 5.0.13a server instances installed into two partitions. There are multiple NICs available and I have allocated two NICs per server instance and they are not used for any other purpose.

Server A: the following config in notes.ini
TCPIP=TCP, 0, 15, 0
TCPIP_TcpIpAddress=0,192.168.151.241:1352
CLUSTER=TCP, 0, 15, 0
CLUSTER_TcpIpAddress=0,172.31.254.252:1352
Ports=TCPIP,CLUSTER
Server_Cluster_Default_Port=CLUSTER
SERVER_CLUSTER_ON=1

Server B: has the following config in notes.ini
TCPIP=TCP, 0, 15, 0
TCPIP_TcpIpAddress=0,192.168.151.242:1352
CLUSTER=TCP, 0, 15, 0
CLUSTER_TcpIpAddress=0,172.31.254.249:1352
Server_Cluster_Default_Port=CLUSTER
Ports=TCPIP,CLUSTER
SERVER_CLUSTER_ON=1

So there you see two servers that each have two NICs with two different subnets. The two NICs that are to be used for cluster communications and probes are connected to each oter via a cross over cable and the connection is good :

Jun 30 21:27:13 jupiter qfe: [ID 517869 kern.info] SUNW,qfe7: 100 Mbps full duplex link up - internal transceiver
Jun 30 21:27:13 jupiter qfe: [ID 517869 kern.info] SUNW,qfe11: 100 Mbps full duplex link up - internal transceiver

I can specify that an issue an ICMP ping out via one interface and arrive at the other and vice versa and this works as expected :

# ping -v -n -i 172.31.254.249 172.31.254.252
172.31.254.252 is alive
# ping -v -n -i 172.31.254.252 172.31.254.249
172.31.254.249 is alive

However both of these nodes seem to refuse to talk with each other as cluster nodes :

When the Domino servers are running I see this :

SERVER A:

> sh po TCPIP
TCP/IP Port Driver
Transport Provider: TCP

Notes Session Local Address Foreign Address
14940005 192.168.151.241:1352 *:*
> sh po CLUSTER
TCP/IP Port Driver
Transport Provider: TCP

Notes Session Local Address Foreign Address
14940006 172.31.254.252:1352 172.31.254.249:34793
14950002 172.31.254.252:1352 172.31.254.252:34794
14990001 172.31.254.252:34794 172.31.254.252:1352
14970003 172.31.254.252:1352 *:*

> sh cluster
Cluster Information
Cluster name: LBA, Server name: kronos/Finance
Server cluster probe timeout: 1 minute(s)
Server cluster probe count: 8
Server cluster default port: CLUSTER
Server availability threshold: 95
Server availability index: 100 (state: AVAILABLE)
Cluster members (2)...
server: inotes/Finance, availability: UNREACHABLE
server: kronos/Finance, availability: UNREACHABLE

At SERVER B :

> sh po TCPIP
TCP/IP Port Driver
Transport Provider: TCP

Notes Session Local Address Foreign Address
14730001 192.168.151.242:1352 *:*
> sh po CLUSTER
TCP/IP Port Driver
Transport Provider: TCP

Notes Session Local Address Foreign Address
14760005 172.31.254.249:1352 172.31.254.249:34795
147D0002 172.31.254.249:34795 172.31.254.249:1352
14790003 172.31.254.249:1352 *:*

> sh cluster
Cluster Information
Cluster name: LBA, Server name: kronos/Finance
Server cluster probe timeout: 1 minute(s)
Server cluster probe count: 8
Server cluster default port: CLUSTER
Server availability threshold: 95
Server availability index: 100 (state: AVAILABLE)
Cluster members (2)...
server: inotes/Finance, availability: UNREACHABLE
server: kronos/Finance, availability: UNREACHABLE


What I find most disturbing about this situation is that I can backup the entire Domino data directory and restore it onto a separate machine, move the cross over cable to an interface on that separate machine and everything works as expected.

Also, server A seems to begin to establish a notes session but never completes the process. Show Server details reveals this :

24D9:0079 Database Server Server for session FD2AFA40 on CLUSTER
24D9:007A Database Server Server for session FD2AFA40 on CLUSTER
24D9:007B Database Server Server for session FD39FA40 on TCPIP

Strange hex id numbers assigned to sessions that never seem to proceed.

Any thoughts would be appreciated.

Dennis