I saw this during my sanity test run before cutting 4.8.0, but it wasn't an obvious fix and also doesn't appear to be directly a fault of sos so I didn't hold the release for it.
The processor plugin is timing out for me locally on the turbostat --debug sleep 10 command. Running this manually shows similar-ish behavior. The command runs and does not exit until I hit enter (regardless of how long I wait to do so), at which point it exits "successfully" with an error:
turbostat: get_rapl_counters: failed to read perf_data (8 1): Success
I've not found anything immediately online related to this, so I'm not sure if it's a known fault condition, or if it's something specific to my personal desktop, and for what it's worth I am running a relatively old AMD Ryzen 5 2600.
Is anyone familiar with turbostat and what might cause this behavior, or how to potentially safeguard against it? If not, I'll just slap a timeout on the add_cmd_output call for slightly longer than the 10 second sleep so we don't have to wait the full 5 minute plugin timeout if this gets hit elsewhere.
I saw this during my sanity test run before cutting
4.8.0
, but it wasn't an obvious fix and also doesn't appear to be directly a fault ofsos
so I didn't hold the release for it.The
processor
plugin is timing out for me locally on theturbostat --debug sleep 10
command. Running this manually shows similar-ish behavior. The command runs and does not exit until I hit enter (regardless of how long I wait to do so), at which point it exits "successfully" with an error:turbostat: get_rapl_counters: failed to read perf_data (8 1): Success
I've not found anything immediately online related to this, so I'm not sure if it's a known fault condition, or if it's something specific to my personal desktop, and for what it's worth I am running a relatively old AMD Ryzen 5 2600.
Is anyone familiar with turbostat and what might cause this behavior, or how to potentially safeguard against it? If not, I'll just slap a
timeout
on theadd_cmd_output
call for slightly longer than the 10 second sleep so we don't have to wait the full 5 minute plugin timeout if this gets hit elsewhere.