What is the meaning of these ERROR messages in the CA Service Desk Manager stdlog?
10/30 10:32:00.59 MyServer spelsrvr 9276 ERROR socket_port.c 802 SLUMP api: Couldn't accept() for socket port (WSAEWOULDBLOCK (10035))
10/30 10:32:00.59 MyServer spelsrvr 9276 ERROR api.c 1419 Fastchannel open attempt failed.
The error messages indicate that CA Service Desk Manager failed to generate a Fast Channel connection, because the required port was not available.
CA Service Desk Manager uses a master communication process called "slump" to facilitate communication between all of its processes.
Once communication between a pair of processes is established by the slump process, the processes may operate through a "Fast Channel" connection. Fast Channel allows key processes to communicate directly between each other, and bypass the general slump communication channel. This increases efficiency as it minimizes the number of steps through which the data needs to be routed.
When Fast Channel becomes unavailable, communication reverts to the slump process and Service Desk Manager continues to run. As such, the messages are harmless if they occur infrequently - there is little impact to application performance. If the messages occur often, the cause should be investigated as inter-process communication is sub-optimal. For example, port restrictions enforced by a firewall can cause this message to always occur, leading to performance degradation.
The CA Service Desk Manager r12.5 CA SDM Implementation Guide explains:
Message Dispatcher (sslump_nxd)
Acts as a common bus or message passing system. Components that need to communicate with each other first register with the Message Dispatcher. When a component sends a message, the Message Dispatcher delivers it to those components that have registered to receive that type of message. If two components communicate so much that it would be inefficient to pass the messages through the Message Dispatcher, they create a fast channel between them. You can view a list of registered components using the slstat utility. This component runs on the primary server.
The "CA Incident and Problem Management" green book available from support.ca.com provides additional information. See, for example, the sections "CA Unicenter Service Desk Architecture" and "Distributed Processes."