{"id":34842,"date":"2019-10-31T22:00:43","date_gmt":"2019-10-31T19:00:43","guid":{"rendered":"https:\/\/prohoster.info\/blog\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina\/"},"modified":"2019-10-31T22:00:43","modified_gmt":"2019-10-31T19:00:43","slug":"ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina","status":"publish","type":"post","link":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina","title":{"rendered":"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p><noindex><a rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/rostelecom\/blog\/454126\/\"><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/1594fcccd46a5cd779c58ddc60e34422.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/a><\/noindex><\/p>\n<p>Komponenta ETL datov\u00e9ho skladu se \u010dasto ocit\u00e1 ve st\u00ednu samotn\u00e9ho skladu a je j\u00ed v\u011bnov\u00e1na m\u00e9n\u011b pozornosti ne\u017e hlavn\u00ed datab\u00e1zi nebo frontov\u00fdm komponent\u016fm, BI a reporting. P\u0159itom z pohledu mechaniky pln\u011bn\u00ed skladu daty hraje ETL kl\u00ed\u010dovou roli a vy\u017eaduje od administr\u00e1tor\u016f stejnou pozornost jako ostatn\u00ed komponenty. Jmenuji se Alexander, moment\u00e1ln\u011b spravuji ETL v Rostelecom a v tomto \u010dl\u00e1nku se pokus\u00edm sd\u00edlet n\u011bkter\u00e9 z v\u00fdzev, kter\u00fdm mus\u00ed \u010delit administr\u00e1tor jednoho z nejzn\u00e1m\u011bj\u0161\u00edch ETL syst\u00e9m\u016f ve velk\u00e9m datov\u00e9m skladu spole\u010dnosti Rostelecom.<br \/>\n<noindex><a rel=\"nofollow\" name=\"habracut\"><\/a><\/noindex><br \/>\nPokud jsou mil\u00ed \u010dten\u00e1\u0159i ji\u017e obezn\u00e1meni s na\u0161\u00edm projektem datov\u00e9ho skladu a s produktem Informatica PowerCenter, mohou p\u0159ej\u00edt rovnou k dal\u0161\u00edmu odd\u00edlu.<\/p>\n<p>P\u0159ed n\u011bkolika lety se v Rostelecomu zrodila a za\u010dala se realizovat my\u0161lenka jednotn\u00e9ho firemn\u00edho datov\u00e9ho skladu. \u0158ada sklad\u016f, kter\u00e9 \u0159e\u0161ily jednotliv\u00e9 \u00fakoly, u\u017e byla vytvo\u0159ena, ale po\u010det sc\u00e9n\u00e1\u0159\u016f rostl, n\u00e1klady na \u00fadr\u017ebu tak\u00e9 stoupaly a bylo jasn\u00e9, \u017ee budoucnost pat\u0159\u00ed centralizaci. Architektonicky se jedn\u00e1 o samotn\u00fd sklad, kter\u00fd se skl\u00e1d\u00e1 z n\u011bkolika vrstev, realizovan\u00fdch na Hadoopu a GreenPlum, s pomocn\u00fdmi datab\u00e1zemi, ETL mechanismy a BI.<\/p>\n<p>P\u0159itom kv\u016fli velk\u00e9mu mno\u017estv\u00ed teritor\u00e1ln\u011b rozpt\u00fdlen\u00fdch, r\u016fznorod\u00fdch zdroj\u016f dat byl vytvo\u0159en speci\u00e1ln\u00ed mechanismus pro export dat, jeho\u017e \u010dinnost je \u0159\u00edzena Informatica. V d\u016fsledku toho se bal\u00ed\u010dky s daty dost\u00e1vaj\u00ed do rozhran\u00ed Hadoopu, po \u010dem\u017e za\u010d\u00ednaj\u00ed procesy na\u010d\u00edt\u00e1n\u00ed dat p\u0159es vrstvy skladu do Hadoopu a GreenPlum, kter\u00e9 \u0159\u00edd\u00ed takzvan\u00fd \u0159\u00edd\u00edc\u00ed mechanismus ETL, realizovan\u00fd v Informatica. T\u00edmto zp\u016fsobem je syst\u00e9m Informatica jedn\u00edm z kl\u00ed\u010dov\u00fdch prvk\u016f, kter\u00e9 zaji\u0161\u0165uj\u00ed fungov\u00e1n\u00ed skladu.<\/p>\n<p>O na\u0161em skladu bude podrobn\u011bji \u0159e\u010deno v jednom z n\u00e1sleduj\u00edc\u00edch p\u0159\u00edsp\u011bvk\u016f.<\/p>\n<p>Informatica PowerCenter\/Big Data Management je v sou\u010dasnosti pova\u017eov\u00e1n za vedouc\u00ed software v oblasti n\u00e1stroj\u016f pro integraci dat. Je to produkt americk\u00e9 spole\u010dnosti Informatica, kter\u00e1 pat\u0159\u00ed mezi nejsiln\u011bj\u0161\u00ed hr\u00e1\u010de v ETL (Extract Transform Load), spr\u00e1v\u011b kvality dat, MDM (Master Data Management), ILM (Information Lifecycle Management) a dal\u0161\u00ed.<\/p>\n<p>N\u00e1mi pou\u017e\u00edvan\u00fd PowerCenter je integrovan\u00fd aplika\u010dn\u00ed server Tomcat, na kter\u00e9m b\u011b\u017e\u00ed aplikace Informatica, vykon\u00e1vaj\u00edc\u00ed jej\u00ed slu\u017eby: <\/p>\n<p><b>Dom\u00e9na<\/b>, a v podstat\u011b je to z\u00e1klad pro v\u0161echno ostatn\u00ed, v r\u00e1mci dom\u00e9ny funguj\u00ed slu\u017eby, u\u017eivatel\u00e9 a komponenty GRID.<\/p>\n<p><b>Administrator Console<\/b>, webov\u00fd n\u00e1stroj pro spr\u00e1vu a monitoring, krom\u011b klienta Informatica Developer, je to hlavn\u00ed n\u00e1stroj pro interakci s produktem<\/p>\n<p><b>MRS, Model Repository Service<\/b>, \u00falo\u017ei\u0161t\u011b metadat, p\u0159edstavuje mezivrstvu mezi datab\u00e1z\u00ed, ve kter\u00e9 jsou metadata fyzicky ulo\u017eena, a klientem Informatica Developer, ve kter\u00e9m prob\u00edh\u00e1 v\u00fdvoj. Repozit\u00e1\u0159e ukl\u00e1daj\u00ed jak popis dat, tak dal\u0161\u00ed informace, v\u010detn\u011b pro \u0159adu dal\u0161\u00edch slu\u017eeb Informatica, nap\u0159\u00edklad rozvrhy spu\u0161t\u011bn\u00ed \u00faloh (Schedules) nebo monitorovac\u00ed data, stejn\u011b jako parametrov\u00e9 sady aplikac\u00ed, kter\u00e9 umo\u017e\u0148uj\u00ed pou\u017e\u00edvat stejnou aplikaci pro pr\u00e1ci s r\u016fzn\u00fdmi zdroji a p\u0159\u00edjemci dat.<\/p>\n<p><b>DIS, Data Integration Service<\/b>, to je slu\u017eba, ve kter\u00e9 prob\u00edhaj\u00ed hlavn\u00ed funk\u010dn\u00ed procesy, v n\u00ed\u017e b\u011b\u017e\u00ed aplikace a prob\u00edh\u00e1 vlastn\u00ed spu\u0161t\u011bn\u00ed Workflow (popisy sekvence mappings a jejich interakce) a Mappings (transformace, bloky, ve kter\u00fdch doch\u00e1z\u00ed k samotn\u00fdm p\u0159evod\u016fm a zpracov\u00e1n\u00ed dat).<\/p>\n<p><b>Konfigurace GRID<\/b> \u2013 v podstat\u011b varianta vybudov\u00e1n\u00ed komplexu s vyu\u017eit\u00edm n\u011bkolika server\u016f, kdy je z\u00e1t\u011b\u017e, spou\u0161t\u011bn\u00e1 DIS, rozlo\u017eena mezi nody (tedy servery, kter\u00e9 jsou sou\u010d\u00e1st\u00ed dom\u00e9ny). V p\u0159\u00edpad\u011b t\u00e9to varianty, krom\u011b rozlo\u017een\u00ed z\u00e1t\u011b\u017ee DIS p\u0159es dodate\u010dnou vrstvu abstrakce GRID, kter\u00e1 spojuje n\u011bkolik nod\u016f, na kter\u00fdch DIS pracuje m\u00edsto na konkr\u00e9tn\u00ed jedn\u00e9 nod\u011b, mohou b\u00fdt tak\u00e9 vytvo\u0159eny dodate\u010dn\u00e9 z\u00e1lo\u017en\u00ed instance MRS. M\u016f\u017eete dokonce implementovat vysokou dostupnost, kdy mohou b\u00fdt extern\u00ed \u017e\u00e1dosti prov\u00e1d\u011bny prost\u0159ednictv\u00edm z\u00e1lo\u017en\u00edch nod\u016f v p\u0159\u00edpad\u011b selh\u00e1n\u00ed hlavn\u00edho. Tento typ konstrukce jsme zat\u00edm odm\u00edtli.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/3829509c50c42ac919cb551c72cbc299.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Informatica PowerCenter, schematicky<\/i><\/p>\n<p>Na po\u010d\u00e1tku pr\u00e1ce v r\u00e1mci dodavatelsk\u00e9ho \u0159et\u011bzce dat se pravideln\u011b objevovaly probl\u00e9my, \u010d\u00e1st z nich d\u00edky v t\u00e9 dob\u011b nestabiln\u00ed pr\u00e1ci Informatica. Chci se pod\u011blit o n\u011bkter\u00e9 z nezapomenuteln\u00fdch okam\u017eik\u016f t\u00e9to s\u00e1gy \u2013 osvojen\u00ed Informatica 10.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/f6247d9b3d73ebdc482b545c5b96ce99.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>P\u016fvodn\u00ed logo Informatica<\/i><\/p>\n<p>Do odpov\u011bdnosti na\u0161eho odd\u011blen\u00ed pat\u0159\u00ed tak\u00e9 dal\u0161\u00ed prost\u0159ed\u00ed Informatica, kde existuje specifika kv\u016fli jin\u00e9 z\u00e1t\u011b\u017ei. Prozat\u00edm si v\u0161ak vzpomenu p\u0159edev\u0161\u00edm na to, jak se Informatica vyv\u00edjela jako ETL komponenta samotn\u00e9ho datov\u00e9ho skladu.<\/p>\n<h2>Jak k tomu do\u0161lo<\/h2>\n<p>\nV roce 2016, kdy jsme za\u010dali odpov\u00eddat za pr\u00e1ci s Informatica, ji\u017e dos\u00e1hla verze 10.0, a pro optimisticky nalad\u011bn\u00e9 kolegy, kte\u0159\u00ed se rozhodovali o nasazen\u00ed produktu s minoritn\u00ed verz\u00ed .0, se v\u0161e zd\u00e1lo jasn\u00e9 \u2014 m\u011bla se pou\u017e\u00edt nov\u00e1 verze! Z pohledu hardwarov\u00fdch zdroj\u016f bylo v\u0161e v t\u00e9 dob\u011b skv\u011bl\u00e9.<\/p>\n<p>Od jara 2016 zaji\u0161\u0165oval provoz Informatica dodavatel, a podle slov n\u011bkolika u\u017eivatel\u016f syst\u00e9mu 'fungovala p\u00e1rkr\u00e1t t\u00fddn\u011b'. Je t\u0159eba vysv\u011btlit, \u017ee datov\u00e9 skladi\u0161t\u011b bylo de facto v f\u00e1zi PoC, v t\u00fdmu nebyli administr\u00e1to\u0159i a syst\u00e9m neust\u00e1le padal z r\u016fzn\u00fdch d\u016fvod\u016f, p\u0159i\u010dem\u017e in\u017een\u00fdr dodavatele ho musel opakovan\u011b restartovat.<\/p>\n<p>Na podzim se v t\u00fdmu objevili t\u0159i administr\u00e1to\u0159i, kte\u0159\u00ed si rozd\u011blili oblasti odpov\u011bdnosti a za\u010dala se vytv\u00e1\u0159et norm\u00e1ln\u00ed pr\u00e1ce na zaji\u0161t\u011bn\u00ed provozu syst\u00e9m\u016f v projektu, v\u010detn\u011b Informatica. Je pot\u0159eba zm\u00ednit, \u017ee tento produkt nen\u00ed \u0161iroce roz\u0161\u00ed\u0159en\u00fd a nem\u00e1 velkou komunitu, kde by bylo mo\u017en\u00e9 naj\u00edt odpov\u011bdi na jak\u00e9koli ot\u00e1zky \u010di vy\u0159e\u0161it jak\u00fdkoli probl\u00e9m. Proto byla velmi d\u016fle\u017eit\u00e1 plnohodnotn\u00e1 technick\u00e1 podpora od rusk\u00e9ho partnera Informatica, d\u00edky kter\u00e9 jsme opravovali v\u0161echny na\u0161e chyby a chyby mlad\u00e9 tehdy Informatica 10.<\/p>\n<p>Prvn\u00ed, co jsme museli ud\u011blat pro v\u00fdvoj\u00e1\u0159e na\u0161eho t\u00fdmu a dodavatele, bylo stabilizovat samotnou Informatica a dos\u00e1hnout funk\u010dnosti webov\u00e9 konzole pro spr\u00e1vu (Informatica Administrator).<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/ac4342681f655b9cdb5710a54546c649.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Tak se mnohdy setk\u00e1vali v\u00fdvoj\u00e1\u0159i Informatica<\/i><\/p>\n<p>Opustil proces zji\u0161\u0165ov\u00e1n\u00ed p\u0159\u00ed\u010din a hlavn\u00ed p\u0159\u00ed\u010dinou v\u00fdpadk\u016f byla sch\u00e9ma interakce softwaru Informatica s datab\u00e1z\u00ed repozit\u00e1\u0159e, kter\u00e1 se nach\u00e1zela na relativn\u011b vzd\u00e1len\u00e9m serveru z hlediska s\u00ed\u0165ov\u00e9ho prost\u0159ed\u00ed. To vedlo k zpo\u017ed\u011bn\u00edm a naru\u0161ovalo to fungov\u00e1n\u00ed mechanism\u016f zaji\u0161\u0165uj\u00edc\u00edch kontrolu stavu dom\u00e9ny Informatica. Po ur\u010dit\u00e9m lad\u011bn\u00ed datab\u00e1ze a zm\u011bn\u00e1ch parametr\u016f Informatica, kter\u00e9 ji u\u010dinily tolerantn\u011bj\u0161\u00ed k zpo\u017ed\u011bn\u00edm datab\u00e1ze, a nakonec aktualizaci verze Informatica na 10.1 a p\u0159enos datab\u00e1ze z p\u0159edchoz\u00edho serveru na server bl\u00ed\u017ee k Informatica, probl\u00e9m p\u0159estal b\u00fdt aktu\u00e1ln\u00ed, a od t\u00e9 doby jsme nezaznamenali podobn\u00e9 v\u00fdpadky.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/e264be8e5336fad067eada9c762b2fc9.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Jedn\u00edm z pokus\u016f o zaji\u0161t\u011bn\u00ed funkce Informatica Monitor<\/i><\/p>\n<p>Situace s administra\u010dn\u00ed konzol\u00ed byla tak\u00e9 kritick\u00e1. Jeliko\u017e prob\u00edhala aktivn\u00ed v\u00fdvoj na podm\u00edn\u011bn\u011b produktivn\u00edm prost\u0159ed\u00ed, kolegov\u00e9 \u010dasto museli analyzovat fungov\u00e1n\u00ed mappings a workflow \"za chodu\". V nov\u00e9 Informatice, v Data Integration Service nen\u00ed samostatn\u00fd n\u00e1stroj pro takov\u00e9 sledov\u00e1n\u00ed, ale v webov\u00e9 administrativn\u00ed konzoli se objevila sekce monitorov\u00e1n\u00ed (Informatica Administrator Monitor), kde lze sledovat fungov\u00e1n\u00ed aplikac\u00ed, workflow a mappings, spu\u0161t\u011bn\u00ed, logy. Ob\u010das se konzole stala zcela nedostupnou, nebo se p\u0159estaly aktualizovat informace o aktu\u00e1ln\u00edch procesech v DIS, nebo se objevily chyby p\u0159i na\u010d\u00edt\u00e1n\u00ed str\u00e1nek.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/1d585c8447275cfaaf4ecbaceb52957b.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Nastaven\u00ed parametr\u016f java pro stabilizaci pr\u00e1ce<\/i><\/p>\n<p>\u0158e\u0161en\u00ed probl\u00e9mu prob\u00edhalo mnoha zp\u016fsoby, prov\u00e1d\u011bly se experimenty se zm\u011bnou parametr\u016f, shroma\u017e\u010fovaly se logy, jstack, pos\u00edlaly se do podpory, sou\u010dasn\u011b prob\u00edhalo aktivn\u00ed googlen\u00ed a prost\u011b se to sledovalo.<\/p>\n<p>Na prvn\u00ed m\u00edst\u011b byla vytvo\u0159ena samostatn\u00e1 MRS pro monitorov\u00e1n\u00ed, jak se pozd\u011bji uk\u00e1zalo, \u017ee to je jeden z hlavn\u00edch spot\u0159ebitel\u016f zdroj\u016f v na\u0161ich prost\u0159ed\u00edch, proto\u017ee spu\u0161t\u011bn\u00ed mappings se d\u011bje velmi intenzivn\u011b. Byly zm\u011bn\u011bny parametry t\u00fdkaj\u00edc\u00ed se java heap a \u0159ada dal\u0161\u00edch.<br \/>\nV d\u016fsledku toho se k dal\u0161\u00edmu upgradu Informatica 10.1.1 poda\u0159ilo stabilizovat pr\u00e1ci konzole a monitoru, v\u00fdvoj\u00e1\u0159i za\u010dali pracovat efektivn\u011bji a pravideln\u00e9 procesy se st\u00e1valy st\u00e1le pravideln\u011bj\u0161\u00edmi.<\/p>\n<p>Zaj\u00edmav\u00fdm m\u016f\u017ee b\u00fdt pro\u017eitek interakce mezi v\u00fdvojem a administrac\u00ed. Ot\u00e1zka spole\u010dn\u00e9ho porozum\u011bn\u00ed, jak v\u0161echno funguje, co lze d\u011blat a co nelze, je v\u017edy d\u016fle\u017eit\u00e1 p\u0159i pou\u017e\u00edv\u00e1n\u00ed slo\u017eit\u00fdch syst\u00e9m\u016f. Proto lze sm\u011ble doporu\u010dit nejprve vy\u0161kolit t\u00fdm administr\u00e1tor\u016f, jak by m\u011blo b\u00fdt software spravov\u00e1no, a t\u00fdm v\u00fdvoj\u00e1\u0159\u016f, jak by m\u011bl ps\u00e1t k\u00f3d a navrhovat procesy v syst\u00e9mu, a teprve pot\u00e9 poslat prvn\u00ed a druh\u00e9 pracovat na v\u00fdsledku. To je skute\u010dn\u011b d\u016fle\u017eit\u00e9, kdy\u017e \u010das nen\u00ed neomezen\u00fdm zdrojem. Mnoho probl\u00e9m\u016f lze vy\u0159e\u0161it i n\u00e1hodn\u00fdm zkou\u0161en\u00edm variant, ale n\u011bkdy n\u011bkter\u00e9 vy\u017eaduj\u00ed apriorn\u00ed znalosti \u2014 n\u00e1\u0161 p\u0159\u00edpad potvrzuje d\u016fle\u017eitost pochopen\u00ed t\u00e9to axiomy.<\/p>\n<p>Nap\u0159\u00edklad p\u0159i pokusu o zapnut\u00ed verzov\u00e1n\u00ed v MRS (jak se nakonec uk\u00e1zalo, pot\u0159ebovali jsme jinou verzi SVN), jsme po n\u011bjak\u00e9 dob\u011b s obavami objevili, \u017ee doba restartu syst\u00e9mu se prodlou\u017eila na n\u011bkolik des\u00edtek minut. Po nalezen\u00ed p\u0159\u00ed\u010diny zpo\u017ed\u011bn\u00ed p\u0159i startu a vypnut\u00ed verzov\u00e1n\u00ed jsme to op\u011bt postavili do po\u0159\u00e1dku.<\/p>\n<p>Mezi znateln\u00fdmi p\u0159ek\u00e1\u017ekami spojen\u00fdmi s Informatica si m\u016f\u017eeme vzpomenout na epickou bitvu s rostouc\u00edmi java vl\u00e1kny. V ur\u010dit\u00e9m okam\u017eiku nastal \u010das na replikaci, tedy roz\u0161\u00ed\u0159en\u00ed nastaven\u00fdch proces\u016f na velk\u00e9 mno\u017estv\u00ed syst\u00e9m\u016f zdroj\u016f. P\u0159itom se uk\u00e1zalo, \u017ee ne v\u0161echny procesy v 10.1.1 fungovaly dob\u0159e a po n\u011bjak\u00e9 dob\u011b DIS se stal nefunk\u010dn\u00edm. Byly nalezeny des\u00edtky tis\u00edc vl\u00e1ken, jejich po\u010det zvl\u00e1\u0161\u0165 v\u00fdrazn\u011b rostl p\u0159i procesu nasazen\u00ed aplikac\u00ed. N\u011bkdy bylo nutn\u00e9 restartovat n\u011bkolikr\u00e1t denn\u011b pro obnoven\u00ed funk\u010dnosti.<\/p>\n<p>Tady je t\u0159eba pod\u011bkovat podpo\u0159e, relativn\u011b rychle byly probl\u00e9my lokalizov\u00e1ny a opraveny pomoc\u00ed EBF (Emergency Bug Fix) \u2013 po tom v\u0161em v\u0161ichni m\u011bli pocit, \u017ee n\u00e1stroj skute\u010dn\u011b funguje.<\/p>\n<h2>Ono to p\u0159ece jen funguje!<\/h2>\n<p>\nK za\u010d\u00e1tku pr\u00e1ce v c\u00edlov\u00e9m re\u017eimu vypadala Informatica n\u00e1sledovn\u011b. Verze Informatica 10.1.1HF1 (HF1 to je HotFix1, dodavatelsk\u00e1 sestava z komplexu EBF) s dodate\u010dn\u011b nainstalovan\u00fdmi EBF, kter\u00e9 opravovaly na\u0161e probl\u00e9my se \u0161k\u00e1lov\u00e1n\u00edm a n\u011bkter\u00e9 dal\u0161\u00ed na jednom serveru ze t\u0159\u00ed, kter\u00e9 tvo\u0159ily GRID, 20 jader x86_64 a \u00falo\u017ei\u0161t\u011b, na obrovsk\u00e9m pomal\u00e9m poli lok\u00e1ln\u00edch disk\u016f \u2014 to <a class=\"wpil_keyword_link\" href=\"https:\/\/prohoster.info\/cs\/server\/\"   title=\"konfigurace serveru\" data-wpil-keyword-link=\"linked\"  data-wpil-monitor-id=\"1000\">konfigurace serveru<\/a> pro Hadoop cluster. On another similar server \u2014 Oracle DB, which is used by both the Informatica domain and the ETL management mechanism. All of this is monitored by standard monitoring tools used by the team (Zabbix + Grafana), from two sides \u2014 Informatica itself with its services, and the loading processes going into it. Currently, both performance and stability, excluding external factors, depend on settings that limit the load.<\/p>\n<p>Separately, we can talk about GRID. The environment was built on three nodes with the possibility of load balancing. However, during testing, it was discovered that due to interaction problems between the instances of our applications, such a configuration did not work as expected, and we temporarily decided to abandon this construction scheme by removing two out of three nodes from the domain. The scheme itself remained the same, and now it is indeed a GRID service, but degenerated to a single node.<\/p>\n<p>Right now, there is a complexity related to performance drops during the regular clearing of the monitoring scheme \u2014 during simultaneously running processes in the CHNN and the running cleanup, failures in the ETL management mechanism may occur. This is currently resolved in a \"quick fix\" manner \u2014 by manually clearing the monitoring scheme, losing all its previous data. This is not too critical for production during normal operational activity, but we are still searching for a proper solution.<\/p>\n<p>From this situation arises another problem \u2014 sometimes multiple launches of our management mechanism occur.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/bf612c20dd2d3d9eeb4b5884d9173156.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Multiple launches of the application lead to the breakdown of the mechanism.<\/i><\/p>\n<p>When launched on a schedule during peak load times, such situations may occur that lead to the mechanism breaking. The problem continues to be resolved manually, and we are searching for a permanent solution.<\/p>\n<p>Celkov\u011b lze shrnout, \u017ee p\u0159i vysok\u00e9 z\u00e1t\u011b\u017ei je velmi d\u016fle\u017eit\u00e9 poskytnout adekv\u00e1tn\u00ed zdroje, co\u017e se t\u00fdk\u00e1 jak hardwarov\u00fdch zdroj\u016f pro samotnou Informatica, tak i pro jej\u00ed datab\u00e1zov\u00fd repozit\u00e1\u0159, a tak\u00e9 zajistit pro n\u011b optim\u00e1ln\u00ed nastaven\u00ed. Krom\u011b toho z\u016fst\u00e1v\u00e1 otev\u0159en\u00e1 ot\u00e1zka, kter\u00e1 sch\u00e9ma um\u00edst\u011bn\u00ed datab\u00e1ze je lep\u0161\u00ed \u2014 na samostatn\u00e9m hostiteli, nebo na tom sam\u00e9m, kde b\u011b\u017e\u00ed software Informatica. Na jedn\u00e9 stran\u011b na jednom serveru to vyjde levn\u011bji a p\u0159i kombinaci se t\u00e9m\u011b\u0159 eliminuje mo\u017en\u00fd probl\u00e9m se s\u00ed\u0165ovou interakc\u00ed, na druh\u00e9 stran\u011b \u2014 z\u00e1t\u011b\u017e na hostitele z datab\u00e1ze se dopl\u0148uje z\u00e1t\u011b\u017e\u00ed od Informatica.<\/p>\n<p>Jako v ka\u017ed\u00e9m v\u00e1\u017en\u00e9m produktu, v Informatica se tak\u00e9 objevuj\u00ed kuri\u00f3zn\u00ed momenty.<br \/>\nJednou, kdy\u017e jsem analyzoval n\u011bjakou havarii, jsem si v\u0161iml, \u017ee v log\u00e1ch MRS je podivn\u011b zaznamen\u00e1n \u010das ud\u00e1lost\u00ed.<\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/98cb76478e740e615ebcc8a548d2856c.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>\u010casov\u00fd dualismus v log\u00e1ch MRS \"by design\"<\/i><\/p>\n<p>Uk\u00e1zalo se, \u017ee \u010dasov\u00e9 zna\u010dky jsou zapisov\u00e1ny ve form\u00e1tu 12 hodin, bez ozna\u010den\u00ed AM\/PM, tedy p\u0159ed polednem nebo po n\u011bm. Dokonce byla pod\u00e1na \u017e\u00e1dost ohledn\u011b tohoto a bylo z\u00edsk\u00e1no ofici\u00e1ln\u00ed odpov\u011b\u010f \u2014 takto to bylo navr\u017eeno, v logu MRS jsou zna\u010dky zapisov\u00e1ny pr\u00e1v\u011b v tomto form\u00e1tu. To znamen\u00e1, \u017ee ob\u010das z\u016fst\u00e1v\u00e1 jist\u00e1 intrika ohledn\u011b \u010dasu vzniku n\u011bjak\u00e9 chyby ERROR\u2026<\/p>\n<h2>Usilovat o lep\u0161\u00ed<\/h2>\n<p>\nDnes je Informatica dostate\u010dn\u011b stabiln\u00ed n\u00e1stroj, kter\u00fd je pohodln\u00fd pro administr\u00e1tora a u\u017eivatele, a neoby\u010dejn\u011b siln\u00fd co do sou\u010dasn\u00fdch mo\u017enost\u00ed a potenci\u00e1lu. Funk\u010dn\u011b mnohon\u00e1sobn\u011b p\u0159ekra\u010duje na\u0161e pot\u0159eby a de facto je nyn\u00ed pou\u017e\u00edv\u00e1n v projektu pon\u011bkud netypick\u00fdm zp\u016fsobem. Obt\u00ed\u017ee jsou \u010d\u00e1ste\u010dn\u011b spojen\u00e9 s t\u00edm, jak funguj\u00ed mechanismy \u2014 specifikou je, \u017ee za kr\u00e1tk\u00e9 obdob\u00ed je spu\u0161t\u011bno velk\u00e9 mno\u017estv\u00ed vl\u00e1ken, kter\u00e1 intenzivn\u011b aktualizuj\u00ed parametersety a pracuj\u00ed s datab\u00e1z\u00ed repozit\u00e1\u0159e, p\u0159i\u010dem\u017e hardwarov\u00e9 zdroje serveru jsou t\u00e9m\u011b\u0159 pln\u011b vyu\u017eity na CPU. <\/p>\n<p>Nyn\u00ed jsme se dostali bl\u00edzko k p\u0159echodu na Informatica 10.2.1 nebo 10.2.2, ve kter\u00fdch byly p\u0159epracov\u00e1ny n\u011bkter\u00e9 vnit\u0159n\u00ed mechanismy, a podpora slibuje absenci \u0159ady aktu\u00e1ln\u011b existuj\u00edc\u00edch probl\u00e9m\u016f s v\u00fdkonem a fungov\u00e1n\u00edm. A z pohledu hardwaru se o\u010dek\u00e1vaj\u00ed servery optim\u00e1ln\u00ed konfigurace pro n\u00e1s, s ohledem na rezervu na nejbli\u017e\u0161\u00ed \u010das vzhledem k r\u016fstu a rozvoji \u00falo\u017ei\u0161t\u011b.<\/p>\n<p>Samoz\u0159ejm\u011b, bude t\u0159eba testov\u00e1n\u00ed, kontrola kompatibility a mo\u017en\u00e1 i architektonick\u00e9 zm\u011bny v oblasti HA GRID. Rozvoj v r\u00e1mci Informatica bude pokra\u010dovat, proto\u017ee v kr\u00e1tkodob\u00e9m horizontu nem\u016f\u017eeme nahradit st\u00e1vaj\u00edc\u00ed syst\u00e9m.<br \/>\nA ti, kdo budou pozd\u011bji zodpov\u011bdn\u00ed za tento syst\u00e9m, ur\u010dit\u011b budou schopn\u00ed dos\u00e1hnout po\u017eadovan\u00fdch ukazatel\u016f spolehlivosti a v\u00fdkonu p\u0159edlo\u017een\u00fdch z\u00e1kazn\u00edky.<\/p>\n<p><i>\u010cl\u00e1nek p\u0159ipravila data management t\u00fdm \"Rostelecomu\"<\/i><\/p>\n<p><img decoding=\"async\" alt=\"Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima administr\u00e1tora\" src=\"\/wp-content\/uploads\/2019\/06\/f09a33f522b0909373bef88aad444ea3.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<i>Aktu\u00e1ln\u00ed logo Informatica<\/i><br \/>\n<br \/>Zdroj: <a content=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/rostelecom\/blog\/454126\/\">habr.com<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>ETL-\u043a\u043e\u043c\u043f\u043e\u043d\u0435\u043d\u0442\u0430 \u0445\u0440\u0430\u043d\u0438\u043b\u0438\u0449\u0430 \u0434\u0430\u043d\u043d\u044b\u0445 \u0447\u0430\u0441\u0442\u043e \u043e\u043a\u0430\u0437\u044b\u0432\u0430\u0435\u0442\u0441\u044f \u0432 \u0442\u0435\u043d\u0438 \u0441\u0430\u043c\u043e\u0433\u043e \u0445\u0440\u0430\u043d\u0438\u043b\u0438\u0449\u0430 \u0438 \u0435\u0439 \u0443\u0434\u0435\u043b\u044f\u0435\u0442\u0441\u044f \u043c\u0435\u043d\u044c\u0448\u0435 \u0432\u043d\u0438\u043c\u0430\u043d\u0438\u044f, \u0447\u0435\u043c \u0433\u043b\u0430\u0432\u043d\u043e\u0439 \u0431\u0430\u0437\u0435 \u0434\u0430\u043d\u043d\u044b\u0445 \u0438\u043b\u0438 \u0444\u0440\u043e\u043d\u0442-\u043a\u043e\u043c\u043f\u043e\u043d\u0435\u043d\u0442\u0435, BI, \u0444\u043e\u0440\u043c\u0438\u0440\u043e\u0432\u0430\u043d\u0438\u0438 \u043e\u0442\u0447\u0451\u0442\u043e\u0432. \u041f\u0440\u0438 \u044d\u0442\u043e\u043c \u0441 \u0442\u043e\u0447\u043a\u0438 \u0437\u0440\u0435\u043d\u0438\u044f \u043c\u0435\u0445\u0430\u043d\u0438\u043a\u0438 \u043d\u0430\u043f\u043e\u043b\u043d\u0435\u043d\u0438\u044f \u0445\u0440\u0430\u043d\u0438\u043b\u0438\u0449\u0430 \u0434\u0430\u043d\u043d\u044b\u043c\u0438, ETL \u0438\u0433\u0440\u0430\u0435\u0442 \u043a\u043b\u044e\u0447\u0435\u0432\u0443\u044e \u0440\u043e\u043b\u044c \u0438 \u0442\u0440\u0435\u0431\u0443\u0435\u0442 \u043d\u0435 \u043c\u0435\u043d\u044c\u0448\u0435 \u0432\u043d\u0438\u043c\u0430\u043d\u0438\u044f \u0430\u0434\u043c\u0438\u043d\u0438\u0441\u0442\u0440\u0430\u0442\u043e\u0440\u043e\u0432, \u0447\u0435\u043c \u043e\u0441\u0442\u0430\u043b\u044c\u043d\u044b\u0435 \u043a\u043e\u043c\u043f\u043e\u043d\u0435\u043d\u0442\u044b. \u041c\u0435\u043d\u044f \u0437\u043e\u0432\u0443\u0442 \u0410\u043b\u0435\u043a\u0441\u0430\u043d\u0434\u0440, \u0441\u0435\u0439\u0447\u0430\u0441 \u044f \u0430\u0434\u043c\u0438\u043d\u0438\u0441\u0442\u0440\u0438\u0440\u0443\u044e ETL \u0432 \u0420\u043e\u0441\u0442\u0435\u043b\u0435\u043a\u043e\u043c\u0435, \u0438 [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":26234,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[688],"tags":[],"class_list":["post-34842","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-administrirovanie"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.3 - aioseo.com -->\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Yuri Gagarin\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.3\" \/>\n\t\t<meta property=\"og:locale\" content=\"cs_CZ\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47\u041e\u0442 \u0435\u0436\u0435\u0434\u043d\u0435\u0432\u043d\u044b\u0445 \u0430\u0432\u0430\u0440\u0438\u0439 \u043a \u0441\u0442\u0430\u0431\u0438\u043b\u044c\u043d\u043e\u0441\u0442\u0438: Informatica 10 \u0433\u043b\u0430\u0437\u0430\u043c\u0438 \u0430\u0434\u043c\u0438\u043d\u0430 | ProHoster\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2019-10-31T19:00:43+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2019-10-31T19:00:43+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47Od ka\u017edodenn\u00edch hav\u00e1ri\u00ed ke stabilit\u011b: Informatica 10 o\u010dima admina | ProHoster","description":"","canonical_url":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"cs_CZ","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041e\u0442 \u0435\u0436\u0435\u0434\u043d\u0435\u0432\u043d\u044b\u0445 \u0430\u0432\u0430\u0440\u0438\u0439 \u043a \u0441\u0442\u0430\u0431\u0438\u043b\u044c\u043d\u043e\u0441\u0442\u0438: Informatica 10 \u0433\u043b\u0430\u0437\u0430\u043c\u0438 \u0430\u0434\u043c\u0438\u043d\u0430 | ProHoster","og:url":"https:\/\/prohoster.info\/cs\/blog\/administrirovanie\/ot-ezhednevnyh-avarij-k-stabilnosti-informatica-10-glazami-admina","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2019-10-31T19:00:43+00:00","article:modified_time":"2019-10-31T19:00:43+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"34842","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":"2026-02-08 21:32:10","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-03-01 02:14:05","updated":"2026-02-08 21:32:10","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/34842","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/comments?post=34842"}],"version-history":[{"count":1,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/34842\/revisions"}],"predecessor-version":[{"id":158197,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/34842\/revisions\/158197"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media\/26234"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media?parent=34842"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/categories?post=34842"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/tags?post=34842"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}