ywkaras commented on issue #11749: URL: https://github.com/apache/trafficserver/issues/11749#issuecomment-2327532754
Are all the (likely inlined) calls to ink_get_hrtime() in EThread::execute_regular() in Yahoo ATS 9? It seems iffy that clock_gettime() is avoiding the kernel call/thunk in RedHat 8. Looks like the COARSE clock is needed ( https://docs.redhat.com/en/documentation/red_hat_enterprise_linux_for_real_time/8/html/optimizing_rhel_8_for_real_time_for_low_latency_operation/assembly_application-timestamping_optimizing-rhel8-for-real-time-for-low-latency-operation#assembly_application-timestamping_optimizing-RHEL8-for-real-time-for-low-latency-operation ). But I don't know why this would affect one YCPI proxy and not another. Looks like you can select the (potentially not in userspace) clock type with a config, proxy.config.system_clock . (Definitely a contender for most dubiously useful ATS config variable.) So maybe double check it's consistent between the two proxies running ATS 10? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
