vpp: aliased to /opt/intel/vtune_amplifier/vpp/collector/vpp-collect
Gathering Platform Profiler collection data. Run vpp-collect stop to finish the collection.
Collection results saved in /home/khb/dhcp-10-1-208-197_20190911-1236.tar.gz.
dhcp-10-1-208-197:/home/khb>ls -lt *.gz
-rw-r--r--. 1 khb khb 0 Sep 11 13:11 dhcp-10-1-208-197_20190911-1236.tar.gz
why would this happen? I did drive the SUT to a load of 60 and swap space was minimal ....
"I assume that you ran "vpp-collect config" first?"
Yes, perhaps I should have been more explicit ... I've had several working sessions before and after. It was only one session which exhibited this behavior. I did drive the system to memory exhaustion ... load average of over 60 (older desktop/server, haswell era supermicro).
"Which VTune version is installed?"
Hopefully someone from the VTune team will chime in....
(Note to Intel folks: Keith Bierman is the only person I have ever known with a job title of "Rabble Rowser" on his business card, so he is clearly a good guy!)
In case any intel folks ARE reading, a question. the various files in the compressed tarfile ... is their format "stable" (that is, if we write scripts which import the results into our toolchain, can we expect them not to break? Is there a specification with version numbering tucked away somewhere?).
The 0-length collection looks like a bug. Do you have a way to reproduce this behavior?
With regard to your question about the data file format - these are generally stable but not fully versioned yet. Specifically
- The version of the EMON data (hardware based telemetry) is stored in emon_sysinfo, e.g. ("version", "V11.10 (public)")
- The version of the OS telemetry is stored in storage.dat, e.g. "Dstat 0.7.2 CSV output"