Entry

Howto add timestamps to log files, i.e., watcher diag statements?

May 22nd, 2001 11:41
Engineer TCSI,

We do now have timestamping in log files when process starts, which
helps to know when a restart of a process occurred.  Each time a
process starts, the following is added to the log file in
$DeployRoot/log/<process>.xlog:
-------------------------------------------
Started : Mon May 21 12:37:31 2001 PDT 
-------------------------------------------
OSP CR1.2 (middle-ware platform for Catalant 2.2) added a feature to 
watcher diag statements. 
Now you can say:
    diag output { debug | timing } =
        {none | pkgFtr | fileLine | timeStamp | all}
                                    ^^^^^^^^^
"timeStamp" indicates that all diagnostic messages should contain  a 
time stamp (only), while "all" now includes a time stamp along with 
file, line #, package and feature. The default is "all".  Example:
  AlarmMgr> 01/05/22 10:54:13.382, fmalAlarmSummary.C(49), Feature
  fmal:fmalAlarmSummary : Constructor (valMoOid &,
  fmalPerceivedSeverity, int) <<
If you want a time stamp in uCORE_ERROR messages, then there is 
currently no timestamp being generated. uCORE_ERROR is what OSP uses to 
report conditions that cannot otherwise be easily conveyed back to the 
user.
To sum up, once Customer has moved to Catalant 2.2, they will get time 
stamps in uDEBUG, but there is not yet any plan for time stamps in 
uCORE_ERROR.
There is also no timestamp on messages that come from Python code, such 
as "print xxxx".