Why does EMS report the message "CPUCOL: write to datafile failed. File error 45"
When starting PROGNOSIS the following is being logged to EMS:
09-04-19 03:27:31 \Node5.$XM07 *TANDEM.MEASURE.H01 004016 Measurement 1
$SYSV02.PRGN850.CPUCOLnn: write to datafile failed. File error 45.
You can safely ignore this error. It is due to the PROGNOSIS implementation for making efficient function calls. The following paragraph describes why this behaviour is also seen for the file called DCOLMA16 as the result of the way this is implemented in STDC, and the same explanation is also valid for the CPU collector:
When the PROGNOSIS STDC object collects Measure data for disk statistics, much of the information can be obtained from system function calls, rather than data collected by Measure. However to have the system function calls available to STDC, Measure needs to be running. The I/O associated with Measure running is high so PROGNOSIS has found a way to ensure that the Measure system function calls are available without having the Measure overheads.
For example: PROGNOSIS starts a DCOLMA16 measurement which will fill up the DCOLMA16 file after a single interval of collection. At this time an error 45 message is generated, to indicate that the Measure file is full and no more data are written to this file. However Measure continues to run, making the system function calls still available. Now PROGNOSIS can gather data on disks without the I/O overhead of a Measure process running. The only cost is the Error 45 message.
Therefore the customer should not be concerned when they see the error 45 message occurring for DCOLMA16, as this is an efficiency design for the product & expected behaviour, not a file size limitation.
