Global Parameters relating to logfiles

The following Database sp083_system_parameters apply to the FRU Process, the API Agent and the Gantt Router:

Parameter

Description

glob_trace_cycle_enabled

Application log files can take one of two forms - cycling, or complete. This parameter determines which is in play for this particular instance.

SQL for changing the way logfiles are managed (cycle or not)

update sp083_system_parameters.value = <0/1> where name = glob_trace_cycle_enabled;

glob_trace_mask_no

Trace mask defining what diagnostic output is produced by server components. This must reference an entry in sp412_trace_masks. This value can be overridden by a component-specific value in sp400_components.

SQL for identifying the current logging level

select component_id, sp400.trace_mask_no, sp412.description from sp400, sp412
where sp400.trace_mask_no = sp412.trace_mask_no(+);

SQL for finding the default if trace_mask_no is NULL

Select sp083.value, sp412.description from sp083, sp412
where sp083.value = sp412.trace_mask_no
and name = ‘glob_trace_mask_no’;

glob_trace_max_file_size

Maximum size of diagnostic log files produced by server components, in Mb. The behaviour when the file size is reached is dependent on the setting of glob_trace_cycle_enabled. A true value means that the file will cycle, with the new entries overwriting some of the older entries.

In either form the size of the log file is limited by this value or, if set, the value configured in the max_log_file_size field of sp400_components.

SQL for finding the default size if max_log_file_size is NULL

Select sp083_system_parameters.value from sp083_system_parameters where name = ‘glob_trace_max_file_size’;

SQL for finding the current log file sizes

Select component_id, max_log_file_size from sp400_components;

It is important to get a reasonable amount of history in the log files (as close to the whole day as possible, especially for the API and FRU); however, setting the size too big can make the log files unmanageable. An initial recommendation is to set the FRU log files to be 100M and the APIAgent and GanttRouter to be 200M.

Once you have set a log file size it is important that you have enough disk space to store all the log files and the archive that you intend keeping. If ServiceOptimizer cannot write to a log file because of insufficient space, then the system will crash.

Optimizer logfile

Size limitations and logging level settings do not apply to these files. Additionally, Optimizer logfiles do not cycle.

Whereabouts

Log files are created in <SPHOME>/<instance>/log where

SPHOME	WINDOWS ONLY: %SPHOME%
instance	the username of the OS account under which the process is being run

Naming

In either mode of working, there will be at least two files created for each process. The first is a 'header' file that will include all the start up information for that process and will have a name of the form <filename>.h.

If glob_trace_cycle_enabled is set true this means that once the size of the log file reaches this limit, it "wraps around" and starts writing again from the start of the file. So only 1 further file will be created.

If glob_trace_cycle_enabled is set false

this means that, once the size of the log file reaches its limit, the file is closed and a new file is opened. The naming convention used for each new file is <filename>.1, <filename>.2 etc

This second form of logging has been provided so that it is possible to work in a mode where no logging is lost, which can be helpful when Support are trying to identify the cause of a specific error. However, it should be noted that creating new files will mean that filestore will be used up at a fairly rapid rate and hence some form of monitoring and archiving of the files off to another area may be required to prevent the system running out of space.

Content and Trace Levels

All entries are timestamped

The timestamp used is the time according to the host machine on which the process is running.
Timestamps are also used to form some file names. The timestamp used is the time according to the host machine on which the master system monitor is running.
Timestamps are always expressed in the SP default timezone (reference sp083_system_parameters value 'tz_Default') (see Time Zones).

Differing levels of information may be written to the log (also known as the "trace level"), and these levels may be set on a per component basis. The default level is held in the Database sp083_system_parameters table. Setting Trace Levels gives further details.

The levels of information can be configured at runtime by ServicePower personnel but this facility is not available as a third-party Support function.

We ask that you do not alter the default set ups without first checking with ServicePower support.

If it is requested that you turn up the level to produce more logging, then this should be done only for the period that is required and, once diagnosis is complete, the level is switched back down again. This is important so that file store size issues do not cause problems.

If you have any concerns about the level that you run with, please contact ServicePower support.

When ServiceOptimizer is first set up, the logging level is set to what is considered a ‘default’ level (up to ‘DEBUG1’). For approximately 95% of the time this level is good. However, for detailed call information, the ‘SOAP’ level tracing may be switched on for the APIAgent.

Browser not supported