Re: parallel query wait events

hopehope_123 wrote:
> Hi ,
>
> When i query a table by using the spss client , the query runs deadly
> slow:
>
> SELECT t0.arac_id c0, t0.kaynak_arac_kod c1, t0.marka_id c2,
> t0.model_id c3,
> t0.unite_no c4, t0.motor_no c5, t0.sasi_no c6, t0.versiyon_no
> c7,
> t0.model_yil c8, t0.arac_km c9, t0.baslangic_tarih c10,
> t0.bitis_tarih c11, t0.kaynak_sistem c12, t0.yukleme_ett_tarih
> c13,
> t0.gunleme_ett_tarih c14, t0.yukleme_tarih c15, t0.gunleme_tarih
> c16,
> t0.ilk_satis_musteri_id c17, t0.ilk_satis_tarih c18,
> t0.ilk_satis_bayi_kod c19, t0.son_servis_musteri_id c20,
> t0.son_servis_tarih c21, t0.son_servis_servis_kod c22,
> t0.son_musteri_id c23, t0.son_kontak_tarihi c24,
> t0.son_kaynak_sistem c25, t0.arac_servis_sayi c26,
> t0.ilk_satis_bayi_il_kod c27
> FROM tofas.tfs_arac t0
>
>
> select * from gv$session_Wait where sid in
> (select sid from gv$px_session where qcsid=102) order by event
>
>
> INST_ID SID SEQ# EVENT P1TEXT P1 P1RAW P2TEXT P2 P2RAW P3TEXT P3 P3RAW WAIT_TIME SECONDS_IN_WAIT STATE
> 2 31 521 PX Deq Credit: send
> blkd sleeptime/senderid 268632063 000000001002FFFF passes 12 000000000000000C qref 4398911479144 00000400338E5D68 0 4 WAITING
> 2 40 501 PX Deq Credit: send
> blkd sleeptime/senderid 268632063 000000001002FFFF passes 12 000000000000000C qref 4398911400208 00000400338D2910 0 4 WAITING
> 2 84 541 PX Deq Credit: send
> blkd sleeptime/senderid 268632063 000000001002FFFF passes 12 000000000000000C qref 4398901629736 0000040032F81328 0 4 WAITING
> 2 103 499 PX Deq Credit: send
> blkd sleeptime/senderid 268632063 000000001002FFFF passes 12 000000000000000C qref 4398901628552 0000040032F80E88 0 4 WAITING
> 2 102 1439 SQL*Net message from client driver
> id 1413697536 0000000054435000 #bytes 1 0000000000000001 0 00 0 23 WAITING
>
> select * from gv$sess_io where sid in
> (select sid from gv$px_session where qcsid=102)
>
>
>
> INST_ID SID BLOCK_GETS CONSISTENT_GETS PHYSICAL_READS BLOCK_CHANGES CONSISTENT_CHANGES
> 2 31 0 535 641 0 0
> 2 40 0 787 899 0 0
> 2 84 0 72 127 0 0
> 2 102 6 31728 629 8 0
> 2 103 0 269 380 0 0
>
> It seems that the query does not request io but waits for the parallel
> query events. What can be the problem?
>
> Kind Regards,
> tolga

Building on Sybrand's comments ... if one FTS is slow ... why would you think more of them would be faster? I would suggest that your turn off PQ and go back to basics ... explain plan, autotrace, and basic tuning with indexes.

Also: What edition and version of Oracle running on what platform and operating system?

-- 
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)