SYMPTOMS
SNA Server 2.11 Service Pack 1 includes support for distributed link
services, where a distributed link service is configured to communicate
over a native LAN session to a link service running on a remote SNA Server,
over TCP/IP sockets, Netware IPX/SPX, or Banyan Vines.
If the distributed link service (SNAREM1, SNAREM2, etc) loses its LAN
session to the remote link service (for example, due to a network outage),
the following Windows NT application event log entry is logged by the
distributed link service:
Event: 23
Source: SNA Server
Description: Connection TOKEN1 using Link Service SNAREM1 failed
Qualifier = 000A
Once this occurs, the connection will remain "Inactive" and SNA Server
will not reconnect to a remote link service on another server (assuming
multiple remote link services were configured in the distributed link
service setup dialog). The administrator will be unable to manually
restart the connection unless the SNA Server service is stopped and
restarted.
NOTE: When connecting over TCP/IP sockets between the distributed link
service and the remote link, it will take over 2 minutes for a network
outage to be detected. The TCP/IP socket "keep alive" interval used to
detect network outages on SNA Server is controlled by the SNA Server
WatchDogTimeout parameter (which defaults to 60 seconds), documented in
the following Microsoft Knowledge Base article:
150511 Preventing SNA Server Keepalives on Idle LAN Connections
STATUS
Microsoft has confirmed this to be a problem in SNA Server version 2.11
Service Pack 1.
This problem was corrected in the latest Microsoft SNA Server 2.11 U.S.
Service Pack. For information on obtaining the service pack, query on the
following word in the Microsoft Knowledge Base (without the spaces):