Hi,
As we understand the questions below, it seems as though you are
running a tool that collects data from the VXML Server activity logs
before the logs have become unlocked. May we suggest to avoid the
below scenarios, you change the way you collect the data from the files
such that you will only open the files when they become unlocked. This
would gaurantee that the calls are completed and ensure the collection
of a full dataset.
Here are answers to some of your questions:
2. Assume log files
are cut by hour. If a call arrives at 9:58 and completes at 10:02, which log
file the call will be recorded, 9-9:59 one or 10-10:59 one?
It will be logged in the 9:59 file.
3. Assume log files are cut by hour. If a call arrives at 9:58 and maxes out the 50K cache at 9:59,
then the first chunk of the call records are written out to the 9-9:59 log
file, and the call and logging continue. At 10:02, the call completes, which
log file the second chunk of call records will be recorded, 9-9:59 one or
10-10:59 one?
The chances of the cache maxing out is slim because the cache is
emptied on a per-call basis. Further testing will be needed to
determine the behavior of the VXML Server in this corner case. Is this
something you can test?
4. Assume log files are cut by hour. If a call arrives at 9:58 for Appl1 and transfer to Appl2 at
9:59 and completes at 10:02, the first part of the call records will be
written to Appl1 9-9:59 log file, but which log file the second part of call
records will be recorded, Appl2 9-9:59 one or Appl2 10-10:59 one?
We believe that the application transfers are treated as two seperate
calls for logging purposes. So the call will be written in two
seperate files. The second half being written in the 10-10:59 file.
This behavior will need to be confirmed through our testing.
I have attached the User Guide for the VXML Server which describes logging in detail.
Regards,
-Chris