Me kēia pou pōkole makemake wau e hoʻopau i kahi kuhi hewa e pili ana i ka nānā ʻana o nā waihona AWR e holo ana ma Oracle Exadata. No nā makahiki he 10, ua kū mau wau i ka nīnau: he aha ka haʻawina o Exadata Software i ka huahana? A i ʻole ka hoʻohana ʻana i nā huaʻōlelo hou: pehea "ka loea" ka hana o kahi waihona kikoʻī?
ʻO ka manawa pinepine kēia nīnau pololei, i koʻu manaʻo, ua pane pololei ʻia me ka pili ʻana i nā helu AWR. Hōʻike ia i ke ʻano kakali ʻōnaehana, e mālama i ka manawa pane e like me ka huina o ka manawa hana o nā kaʻina hana (DB CPUs) a me ka manawa kali o nā papa like ʻole.
Me ka hiki ʻana mai o Exadata, ua ʻike ʻia nā manaʻolana kikoʻī e pili ana i ka hana o Exadata Software i nā helu AWR. Ma ke ʻano maʻamau, hoʻomaka nā inoa o ia mau kali me ka huaʻōlelo "cell" (kapa ʻia ke kikowaena Exadata Storage he cell), ʻo ia ka mea maʻamau e kali me nā inoa wehewehe ponoʻī "cell smart table scan", "cell multiblock". heluhelu kino” a me ka “heluhelu kino kino hookahi pahu hookahi”.
I ka hapanui o nā hihia, he liʻiliʻi ka māhele o ia Exadata i ka manawa pane a no laila ʻaʻole lākou e hāʻule i loko o ka Top10 Foreground Events e ka Total Wait Time section (i kēia hihia, pono ʻoe e ʻimi iā lākou ma ka Foreground Wait. ʻāpana hanana). Me ka paʻakikī nui, ua loaʻa iā mākou kahi laʻana o ka AWR i kēlā me kēia lā mai kā mākou mea kūʻai aku, kahi i hoʻokomo ʻia ai nā manaʻo Exadata i ka ʻāpana Top10 a ʻo ka huina he 5%:
hanana
E kali
Ka nui o ka manawa kali (sec)
Awg Kali
%DB manawa
E kali Papa
DB CPU
115.2K
70.4
SQL*Net nā ʻikepili hou aʻe mai dblink
670,196
5471.5
8.16ms
3.3
Leakaʻaʻike
heluhelu kino hookahi poloka
5,661,452
3827.6
676.07us
2.3
Mea hoʻohana I/O
Hoʻopili i ke kaulike ASM
4,350,012
3481.3
800.30us
2.1
Other
Heluhelu kino ke kelepona multiblock
759,885
2252
2.96ms
1.4
Mea hoʻohana I/O
heluhelu pololei
374,368
1811.3
4.84ms
1.1
Mea hoʻohana I/O
SQL*Net memo mai dblink
7,983
1725
216.08ms
1.1
Leakaʻaʻike
paʻa paʻa paʻa paʻa
1,007,520
1260.7
1.25ms
0.8
Mea hoʻohana I/O
ala pololei heluhelu temp
520,211
808.4
1.55ms
0.5
Mea hoʻohana I/O
enq: TM - paio
652
795.8
1220.55ms
0.5
noi
Lawe pinepine ʻia kēia mau hopena mai nā helu AWR:
1. ʻAʻole kiʻekiʻe ka hāʻawi ʻana o Exadata magic i ka hana ʻikepili - ʻaʻole ia ma mua o 5%, a ʻo ka waihona "exadatizes" maikaʻi ʻole.
2. Inā hoʻoili ʻia kēlā waihona mai Exadata i ka papa hana "server + array", a laila ʻaʻole e loli nui ka hana. No ka mea, inā ʻoi aku ka liʻiliʻi o kēia ʻano i ʻekolu manawa ʻoi aku ka lohi ma mua o ka ʻōnaehana mālama Exadata (he mea paʻakikī ʻole no ka All Flash arrays hou), a laila e hoʻonui i ka 5% me ʻekolu e loaʻa iā mākou ka piʻi ʻana o ka māhele o ka I/O kali i 15% - e ola maoli ka waihona i kēia!
ʻAʻole pololei kēia mau hopena ʻelua, ʻo ia hoʻi, hoʻololi lākou i ka ʻike o ka manaʻo ma hope o Exadata Software. ʻAʻole hāʻawi wale ʻo Exadata i ka I/O wikiwiki, hana ʻokoʻa ia i ka hoʻohālikelike ʻana i ke kikowaena kikowaena + array architecture. Inā "exadapted" maoli ka hana ʻikepili, a laila hoʻololi ʻia ka logic SQL i ka ʻōnaehana mālama. ʻO nā kikowaena mālama, mahalo i kekahi mau mīkini kūikawā (ʻo Exadata Storage Indexes, akā ʻaʻole wale), e ʻimi i ka ʻikepili pono iā lākou iho a hoʻouna i ka DB i nā kikowaena. Hana maikaʻi lākou i kēia, no laila he liʻiliʻi ka māhele o Exadata maʻamau i ka manawa pane.
Pehea e loli ai kēia māhele ma waho o Exadata? Pehea e pili ai kēia i ka hana o ka ʻikepili holoʻokoʻa? ʻO ka hoʻāʻo ʻana e pane maikaʻi i kēia mau nīnau. No ka laʻana, e kali ana i kahi "cell smart table scan" ma waho o Exadata hiki ke hoʻololi i kahi papaʻaina piha piha e lawe ai ʻo I/O i ka manawa pane holoʻokoʻa a haʻahaʻa loa ka hana. ʻO ia ke kumu he hewa, i ka nānā ʻana iā AWR, e noʻonoʻo i ka huina o nā manaʻo Exadata e like me ka hāʻawi ʻana o kāna kilokilo i ka hana, a ʻoi aku hoʻi e hoʻohana i kēia pākēneka e wānana i ka hana ma waho o Exadata. No ka hoʻomaopopo ʻana i ka "pololei" o ka hana o ka waihona, pono ʻoe e aʻo i nā helu AWR o ka ʻāpana "Instance Activity Stats" (he nui nā helu me nā inoa wehewehe ponoʻī) a hoʻohālikelike iā lākou me kekahi.
A no ka hoʻomaopopo ʻana i ka manaʻo o kahi waihona ma waho o Exadata, ʻoi aku ka maikaʻi o ka hana ʻana i kahi clone database mai kahi hoʻihoʻi ma ka hoʻolālā ʻana a nānā i ka hana o kēia clone ma lalo o ka ukana. Loaʻa i nā mea nona Exadata, ma ke ʻano he kūlana.
Author: ʻO Alexey Struchenko, ke poʻo o ka ʻoihana ʻikepili Jet Infosystems
Source: www.habr.com