On Wed, Dec 28, 2011 at 1:19 AM, umapathi b <umapathi.b@stripped> wrote:
> Hi All,
> Yesterday suddenly we got tens of slow query mails running on 10.10.90.57
> ( houdb01 ) . When I inspected the show process list output ..there were
> two queries which have been running for the longer time compared to others
> Plz find the attached file for inspection.
> 1) (SELECT UI.USER_ID FROM (user_info UI LEFT OUTER JOIN user_contact UC
> ON UI.USER_ID = UC.USER_ID),course C WHERE UI.COURSE_ID = C.COURSE_ID AND
> UI.DRIVERS_LICENSE LIKE '05105915%' AND C.PRODUCT_ID=25) UNION (SELECT
> UI.USER_ID FROM (user_info UI LEFT OUTER JOIN user_contact UC ON
> UI.USER_ID = UC.USER_ID),course C WHERE UI.COURSE_ID = C.COURSE_ID AND
> UI.LOGIN LIKE '05105915%' AND C.PRODUCT_ID=25)
> 2) UPDATE user_info SET TEST_INFO = '100000' WHERE USER_ID = 18921819
> These queries were running with the same time .i.e. each query started at
> the same time ..and were running parallelly with the same no.of seconds .
> Then I killed the Select Query ( the first one ) assuming that it will
> free other resources . But this didn't happen and the thread status was
> showing as killed but not releasing others ..
> At this time there was not even any load on the server . In this regard ,
> I re-started mysqld on 57 which resolved the issue . Infact , the queries
> mentioned above were not even taking any time to be executed and they were
> not slow queries .
> I inspected all the management,data node,mysql logs but could not find any
> error/information wrt to this ..
> Can anybody clarify this , why this happened so ?
> What measures are to be taken so that this kind of things are not repeated again ..