OSI/FTAM Configuration and Management Manual

Planning, Installing, and Configuring Tandem FTAM
OSI/FTAM Configuration and Management Manual—421944-001
3-36
Monitoring Protocol Errors
Resource Allocation Interdependencies
Obviously, there is some overlap between these functional groupings. For example,
where you allocate resources such as code files, swap volumes, and the MIB file affects
how your system performs. Keep in mind that several of the attributes discussed here are
interdependent.
Consider how the setting for one attribute might affect other attributes. Note the
following examples:
•
If you set a low number for the maximum number of processes you can configure
(using the MAXFTAMI and MAXFTAMR attributes) and also set a low value for
the number of associations each process can have (using the MAXASSOC
attribute), you may be unable to add enough additional processes to handle a heavy
workload.
•
The value of the MAXASSOC attribute for a responder process is the upper limit for
the value of the MAXATTACH attribute. Therefore, if you decrease the value of
MAXASSOC, the MAXATTACH attribute might also need to be decreased. Under
most circumstances, you should set the two attributes to the same value.
•
Specifying as the swap volume a disk that is not heavily used and has significant
available space can improve the performance of your Tandem FTAM processes.
Monitoring Protocol Errors
One way to monitor the activity of your system is to keep track of the number of
protocol errors a process is detecting. You can use the THRESHOLD attribute of the
PROCESS object to modify the frequency at which these protocol errors are reported.
The value of this attribute determines the number of protocol errors a process detects
before it sends an event message.
Typically, an initiator or responder process detects a protocol error because the remote
entity appears to be violating the protocol in ISO 8571—for example, by using a service
class or attribute that has not been negotiated. Protocol errors are very serious in the user
environment, because they prevent interoperation.
The THRESHOLD attribute has a value in the range of 0 to 32,767 and defaults to 0.
When the value is zero (0), the process does not issue event messages for protocol
errors, regardless of the number of them it detects. When the THRESHOLD attribute is
set to a nonzero value, the protocol-error counter increments as each protocol error is
detected. Whenever the counter reaches any multiple of the threshold value, the process
issues an event message without resetting the counter. For example, if the
THRESHOLD value is set to ten, a message is issued for every ten protocol errors
detected. You can reset the counter to zero by suspending (or stopping) the process and
then activating (or starting) it.