To run a 8 hour and 30 minute test, logging should be fully disabled for these kind of test. According to HPE best practices ''you should disable all logging by ensuring that the Enable logging checkbox in the runtime settings dialog's Log branch is unchecked, unless you need it for debugging purposes (and with no more than 10 Vusers).
Running many Vusers with logging can affect the test results because logging is a time-consuming activity. Each log message also takes up space on the disk, and in extreme circumstances, particularly on long tests, could cause the disk to fill up. You should also disable the Send messages only when an error occurs option, since there may be a lot more errors than expected.''
For the Controller Output messages of 8704 errors in this test. I highly suggest that the two scripts be reviewed and the errors fixed that was received during the test.
You will need to drill down on the % Disk Time for each LG, I positive it will show that maxed out physical disk on the LGs at various points within the test.
When the physical disk runs out of space at points during the test, these file are not create, thus they are going to be missing from the results.
Run a one hour test to ensure that the scripts are functioning as needed before running a 8 hour test. When running a long test, logging should be fully disabled.