NonStop Server Guide for BEA WebLogic Server 8.1

Using the WS Plug-in
HP NonStop Server Guide for BEA WebLogic Server 8.1—529891-001
8-9
Defining the WS Plug-in Server Class
problems. Level two should only be used if instructed to do so by HP support staff.
If you enable tracing, you must also specify the Stderr standard error file.
If you want a separate log file for each WS Plug-in server process, set LOG_DIR to
the pathname of an existing directory. Each WS Plug-in process will create a
separate file in that directory to which tracing information will be written. The WS
Plug-in process name (without the leading dollar sign) is used for the filename.
RETRY_PERIOD
Determines the number of seconds the WS Plug-in servers wait before attempting
to communicate with a previously failed target application server process in a
cluster. This variable applies only to prefixes for which J2EE mode is enabled. The
default value is 30 seconds.
Note that this period is the minimum delay; the actual delay could be almost as
long as RETRY_PERIOD plus STATS_PERIOD because the WS Plug-in processes
only check the cluster status table if the fault indicator in the prefix status table has
been updated since that WS Plug-in process last checked the cluster status table.
If no other server processes fail after a target application server process is marked
as failed, the WS Plug-in processes will not be advised to check the cluster status
table and, therefore, will not try to contact the failed server again. However, the WS
Plug-in checks the cluster status table at the start of each STATS_PERIOD.
STATS_PERIOD
Sets the interval in seconds between recalculations of the capacity of each target
application server process in the cluster. The default is 30 seconds. Note that each
WS Plug-in process collects and reports its own statistics separately. Also,
statistics are recorded and reported separately for each URL prefix configured. A
WS Plug-in process will only check for STATS_PERIOD expiry when it receives a
request for a particular prefix. Thus if no requests for a given prefix are received by
the WS Plug-in process for a period greater than the configured period, the
STATS_PERIOD will be extended to the time between requests.
SHOW_STATS
controls whether capacity and workload statistics are displayed.
ON usage statistics for each server in the cluster are written to
the log file at the end of each STATS_PERIOD
EMS usage statistics are written to the Event Message System
log
Example 8-3. WS Plug-in Statistics Output
Prefix: prefix, Address: addr, Port: port, Capacity: c, New
Sessions: sessions, Posts: posts