Also additional information. Before fix busy_time/cpu_time from THD collected to global tables just after connection end.
My fix collect this information when you execute SHOW [CLIENT|THREAD|USER]_STATS.
This fix doesn't fix this bug directly. For fix of bug you should compile source code with HAVE_CLOCK_GETTIME.
At least my local build compiled with this option.
My tests doesn't include direct check of CPU time workability (this is required additional debug variables/so one for stable test), and probably our packages build without it.
Also additional information. Before fix busy_time/cpu_time from THD collected to global tables just after connection end. THREAD| USER]_STATS.
My fix collect this information when you execute SHOW [CLIENT|
This fix doesn't fix this bug directly. For fix of bug you should compile source code with HAVE_CLOCK_GETTIME.
At least my local build compiled with this option.
My tests doesn't include direct check of CPU time workability (this is required additional debug variables/so one for stable test), and probably our packages build without it.
Can Nacho check this?