|  | Marcel,
   Times are in the .TIMES file ;-) Seriously, for each line of data
   in the .LOG file you will find a corresponding record in the .TIMES
   file giving the time of the log line and its offset from the beginning
   of the .LOG file, all this helps us timestamp the data and search for
   specific times in the monitor interface.
   I STRONGLY urge you to NOT change ANYTHING in any of these three files
   (.TIMES, .LOG and .EVENT) as you will completely screw up PCM and its
   ability to retrieve context etc.
   Are you trying to stop the system with a BREAK SIGNAL or CONTROL-P does
   the system expect the one you are trying to use? I dont have access to
   a 3100-20 so I cant be much more help I am afraid.
   All the text does not appear because ENS gets the text from the log file
   which the daemon may not have updated with the FULL line of context
   because it has not appeared on the console line (this is more likely to
   happen on slow consoles where the event will be detected x characters
   before a newline character appears on the line). Dont forget we generate
   an event IMMEDIATELY after the last character of you event text definition
   has occured. If you want to guarantee that the whole line is available,
   change your event definition to be a regular expression and add the
   Start/End of line metacharacter to the end of the definition e.g.
         BEFORE, text is LITERAL, value is "Username:"
	 This may not show the value entered at the username prompt, to see
         the value, enter your event as follows:
         AFTER, text is REGULAR EXPRESSION, value is "Username:*^"
         This will generate an event when the user has entered their username
         and hit <RETURN>, you action routines will show what they typed.
         (The event will also get generated if loginout times out the login
         prompt.
  Hope this helps.
Cheers,
Phil
 | 
|  |     
>>   Times are in the .TIMES file ;-) Seriously, for each line of data
>>   in the .LOG file you will find a corresponding record in the .TIMES
>>   file giving the time of the log line and its offset from the beginning
>>   of the .LOG file, all this helps us timestamp the data and search for
>>   specific times in the monitor interface.
    
    I assume it has to do with speed too. Searching through .LOG using
    offset is faster than the VCS way of searching from start through end.
>>   Are you trying to stop the system with a BREAK SIGNAL or CONTROL-P does
>>   the system expect the one you are trying to use? I dont have access to
>>   a 3100-20 so I cant be much more help I am afraid.
    
    Tried F5 key as this is in use with a couple of MV systems to go into
    halt mode. But this returns 'Username:'
    Tried ^P, PCM asks 'Do you want to send BREAK', answer 'Y', nothing
    happens.
>>   has occured. If you want to guarantee that the whole line is available,
>>   change your event definition to be a regular expression and add the
>>   Start/End of line metacharacter to the end of the definition e.g.
    
    Thanx for pointer of '^' as EOL character check. Have to get used to
    the new regular expression stuff.
    
    Regards, Marcel.
 | 
|  | Marcel,
>>>>   Times are in the .TIMES file ;-) Seriously, for each line of data
>>>>   in the .LOG file you will find a corresponding record in the .TIMES
>>>>   file giving the time of the log line and its offset from the beginning
>>>>   of the .LOG file, all this helps us timestamp the data and search for
>>>>   specific times in the monitor interface.
    
>>    I assume it has to do with speed too. Searching through .LOG using
>>    offset is faster than the VCS way of searching from start through end.
  Yup, it also gives us portability as we dont have Indexed files on ULTRIX
  or OSF/1
>>>>   Are you trying to stop the system with a BREAK SIGNAL or CONTROL-P does
>>>>   the system expect the one you are trying to use? I dont have access to
>>>>   a 3100-20 so I cant be much more help I am afraid.
    
>>    Tried F5 key as this is in use with a couple of MV systems to go into
>>    halt mode. But this returns 'Username:'
>>    Tried ^P, PCM asks 'Do you want to send BREAK', answer 'Y', nothing
>>    happens.
  F5 on VTxxx terminals actually used to transmit a BREAK SIGNAL, F5 on a
  workstation sends who knows what, and it looks like your MicroVAX wants
  Control-P and not the BREAK SIGNAL.
  So, try a SET BREAK_KEY CTRL_I within the monitor interface, then connect
  to the system and type ^P, this will send a control-P instead of a break
  sytstem
>>>>   has occured. If you want to guarantee that the whole line is available,
>>>>   change your event definition to be a regular expression and add the
>>>>   Start/End of line metacharacter to the end of the definition e.g.
    
>>    Thanx for pointer of '^' as EOL character check. Have to get used to
>>    the new regular expression stuff.
 
  No problem, there are some other "metacharacters" too, I suggest that you
  look them up, they can be very powerful.
Cheers,
Phil
 |