Andrei Elkin wrote:
> DaoGang, hello.
>
>
>
>> Hi Zhenxing and andrei,
>> Please review the following patch:
>> http://lists.mysql.com/commits/115757
>>
>>
>
> I have started to stumble at the following point:
>
> While the bug description blames the relay log's FD timestamp saying
> `The displayed value for
> Seconds_Behind_Master is calculated based on the timestamp of the relay log's
> Format_Description_Log_Event'
> At the same time
>
> if (!(ev->is_artificial_event() || ev->is_relay_log_event() ||
> (ev->when == 0)))
>
> the if argument should not let the slave side FD event to update
>
No. The rli->last_master_timestamp is updated by the slave side FD.
The above 'if argument' really lets the slave side FD event to pass
and update rli->last_master_timestamp by the FD's timestamp.
> {
> - rli->last_master_timestamp= ev->when + (time_t) ev->exec_time;
>
> rli->last_master_timestamp.
>
> + time_t rli_last_master_timestamp= rli->last_master_timestamp;
> + if (ev->get_type_code() != FORMAT_DESCRIPTION_EVENT)
>
The 'rpl_seconds_behind_master.test' will abort if we don't add the above
if argument 'ev->get_type_code() != FORMAT_DESCRIPTION_EVENT'. And
It certifies that the rli->last_master_timestamp is updated by the slave
side FD too.
> + {
> + rli->last_master_timestamp= ev->when + (time_t) ev->exec_time;
> + }
> + DBUG_ASSERT(rli_last_master_timestamp <= rli->last_master_timestamp);
> DBUG_ASSERT(rli->last_master_timestamp >= 0);
> }
>
> I find it puzzling. Could you please explain that, and if necessary to
> alter the patch comments.
>
Now. What puzzles you? What should be updated for patch comments?
Thanks!
Best Regards,
Daogang
> Thanks!
>
> Andrei
>
