Traces shortened by NUM_USECS.I causes hang

[clarkanderson]

There appears to be a bug when truncating the run time via the config file "NUM_USECS.I" parameter. It correctly filters out ops that have timestamps > the parm value. It then writes "Done prepareIOs 167895 requests." to the standard output, but never does the end of run calculations for throughput and response time. It hangs there.

I've only tried this on AIX so far. It may, or may not work in Linux.

[arh]

There might be a bug there. I have never tried to limit execution by num_usec.i option.

…

On Thu, Jan 11, 2018 at 3:29 PM clarkanderson ***@***.***> wrote: There appears to be a bug when truncating the run time via the config file "NUM_USECS.I" parameter. It correctly filters out ops that have timestamps > the parm value. It then writes "Done prepareIOs 167895 requests." to the standard output, but never does the end of run calculations for throughput and response time. It hangs there. I've only tried this on AIX so far. It may, or may not work in Linux. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#6>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAgp38FsSl8l7VOTOp4SyxaGYLUbCLciks5tJn0qgaJpZM4Rbed2> .