> As discussed yesterday on IRC:
>
> - THD::update_server_status uses utime_after_query too
> do you know whether updating the field might cause the change of behavior
> somewhere else?
I have checked the code and did not find side effects now. Moreover I ran tests and everything seems ok: http://jenkins.percona.com/view/PS%205.6/job/percona-server-5.6-param/804/ .
I have changed it for some reason, but unfortunately I don't remember the reason. I tested "utime_after_lock" variant both with regression test and manually with different test cases and it works well so I decided to revert this line.
> As discussed yesterday on IRC: server_ status uses utime_after_query too jenkins. percona. com/view/ PS%205. 6/job/percona- server- 5.6-param/ 804/ .
>
> - THD::update_
> do you know whether updating the field might cause the change of behavior
> somewhere else?
I have checked the code and did not find side effects now. Moreover I ran tests and everything seems ok: http://
> - time_collect( thd->utime_ after_query - after_lock) ; time_collect( query_type, thd->utime_ after_query -
> 64 - query_response_
> thd->utime_
> 65 + query_response_
> thd->start_utime);
> before: utime after lock
> after: start utime
> why?
I have changed it for some reason, but unfortunately I don't remember the reason. I tested "utime_after_lock" variant both with regression test and manually with different test cases and it works well so I decided to revert this line.