{"id":84108,"date":"2020-06-05T07:42:44","date_gmt":"2020-06-05T05:42:44","guid":{"rendered":"https:\/\/prohoster.info\/blog\/administrirovanie\/tushit-li-servera-esli-zagorelsya-smouk-test-dataczentra"},"modified":"2020-06-05T07:42:44","modified_gmt":"2020-06-05T05:42:44","slug":"tushit-li-servera-esli-zagorelsya-smouk-test-dataczentra","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/administrirovanie\/tushit-li-servera-esli-zagorelsya-smouk-test-dataczentra","title":{"rendered":"A duhet t\u00eb \"shuhen\" server\u00ebt kur \u00ebsht\u00eb \"shk\u00ebndij\u00eb\" testi i avullit t\u00eb datacentrit?","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\u00c7far\u00eb do t\u00eb ndihnit n\u00ebse nj\u00eb dit\u00eb t\u00eb bukur vere datacentri me pajisjet tuaja do t\u00eb dukej k\u00ebshtu?<\/p>\n

\"A
\n
\nP\u00ebrsh\u00ebndetje t\u00eb gjith\u00ebve! Emri im \u00ebsht\u00eb Dmitrij Samsonov, un\u00eb punoj si administrator kryesor i sistemeve n\u00eb \"Odnoklassniki<\/a><\/noindex>\". N\u00eb fotografin\u00eb e k\u00ebsaj fotografie \u00ebsht\u00eb nj\u00eb nga kat\u00ebr datacentrat ku ndodhen pajisjet q\u00eb sh\u00ebrbejn\u00eb projektin ton\u00eb. Pas k\u00ebtyre mur\u00ebve ndodhen rreth 4,000 nj\u00ebsi pajisje: server\u00eb, sisteme ruajtjeje t\u00eb t\u00eb dh\u00ebnave, pajisje rrjetesh etj. \u2014 gati \u2153 e gjith\u00eb pajisjeve tona.
\nShumica e server\u00ebve jan\u00eb me Linux. Ka edhe nj\u00eb duzin\u00eb server\u00ebsh me Windows (MS SQL) \u2014 trash\u00ebgimia jon\u00eb, nga e cila kemi filluar t\u00eb heqim dor\u00eb gradualisht p\u00ebr shum\u00eb vite.
\nPra, m\u00eb 5 qershor 2019, n\u00eb ora 14:35, inxhinier\u00ebt e nj\u00ebrit prej datacentrove tona raportuan p\u00ebr nj\u00eb alarm zjarri.
\n<\/a><\/noindex><\/p>\n

Mohim<\/h4>\n

\n14:45. Incidentet e vogla me tym n\u00eb datacentre ndodhin m\u00eb shpesh se sa duket. T\u00eb dh\u00ebnat brenda sallave ishin normale, prandaj reagimi yn\u00eb i par\u00eb ishte relativisht i qet\u00eb: vendos\u00ebm nj\u00eb ndales\u00eb p\u00ebr pun\u00ebt me prodhimin, q\u00eb do t\u00eb thot\u00eb \u00e7do ndryshim konfiguracionesh, publikim t\u00eb versioneve t\u00eb reja etj., p\u00ebrve\u00e7 pun\u00ebve q\u00eb lidhen me riparimin e ndonj\u00eb gj\u00ebje.<\/p>\n

Nervoziz\u00ebm<\/h4>\n

\nA keni provuar ndonj\u00ebher\u00eb t\u00eb pyesni zjarrfik\u00ebsit, n\u00eb cilin vend specifik n\u00eb \u00e7ati ndodhi zjarri, ose t\u00eb shkoni vet\u00eb n\u00eb \u00e7atin e djegur p\u00ebr t\u00eb vler\u00ebsuar situat\u00ebn? Sa do t\u00eb jet\u00eb niveli i besimit n\u00eb informacionin e marr\u00eb p\u00ebrmes pes\u00eb njer\u00ebzve?<\/p>\n

14:50. Ka ardhur informacion q\u00eb zjarri po afrohet deri te sistemi i ftohjes<\/b>. Por, a do t\u00eb arrij\u00eb? Administrator i sistemit n\u00eb detyr\u00eb po nxjerr trafikun e jasht\u00ebm nga frontet e k\u00ebtij datacentri.<\/p>\n

Momentalisht, frontet e t\u00eb gjith\u00eb sh\u00ebrbimeve tona jan\u00eb t\u00eb dyfishuara n\u00eb tre datacentre, me balansim n\u00eb nivelin DNS, e cila lejon heqjen e adresave t\u00eb nj\u00eb datacentri nga DNS, duke k\u00ebshtu mbrojtur p\u00ebrdoruesit nga probleme t\u00eb mundshme me qasjen n\u00eb sh\u00ebrbime. N\u00eb rast se problemet n\u00eb datacent\u00ebr tashm\u00eb kan\u00eb ndodhur, ai automatikisht del nga rotacioni. M\u00eb shum\u00eb mund t\u00eb lexoni k\u00ebtu: Balansimi i ngarkes\u00ebs dhe q\u00ebndrueshm\u00ebria n\u00eb \"Odknoklassniki\".<\/a><\/noindex><\/p><\/blockquote>\n

\nP\u00ebr ne zjarri nuk ka pasur asnj\u00eb ndikim deri tani \u2014 as p\u00ebrdoruesit, as pajisjet nuk kan\u00eb p\u00ebsuar. \u00cbsht\u00eb nj\u00eb fatkeq\u00ebsi? Seksioni i par\u00eb i dokumentit \"Plani i veprimeve n\u00eb rast fatkeq\u00ebsie\" jep nj\u00eb definicion p\u00ebr konceptin \"Fatkeq\u00ebsi\", dhe p\u00ebrfundon seksioni k\u00ebshtu:
\n\u00abN\u00ebse ka dyshime, \u00ebsht\u00eb fatkeq\u00ebsi ose jo, at\u00ebher\u00eb \u00ebsht\u00eb fatkeq\u00ebsi!<\/u>\u00bb<\/b><\/p>\n

14:53. Em\u00ebrohet koordinatori i fatkeq\u00ebsis\u00eb.<\/p>\n

Koordinatori \u00ebsht\u00eb personi q\u00eb kontrollon komunikimin midis t\u00eb gjith\u00eb pjes\u00ebmarr\u00ebsve, vler\u00ebson p\u00ebrmasat e fatkeq\u00ebsis\u00eb, p\u00ebrdor \"Planin e veprimeve n\u00eb rast fatkeq\u00ebsie\", angazhon personelin e nevojsh\u00ebm, kontrollon p\u00ebrfundimin e riparimit, m\u00eb e r\u00ebnd\u00ebsishmja \u2014 delegon \u00e7do detyr\u00eb. N\u00eb fjal\u00eb t\u00eb tjera, ky \u00ebsht\u00eb personi q\u00eb menaxhon t\u00eb gjith\u00eb procesin e likuidimit t\u00eb fatkeq\u00ebsis\u00eb.<\/p><\/blockquote>\n

<\/p>\n

Tregu<\/h4>\n

\n15:01. Fillojm\u00eb t\u00eb \u00e7aktivizojm\u00eb serverat q\u00eb nuk jan\u00eb t\u00eb lidhur me prodhimin.
\n15:03. \u00c7aktivizojm\u00eb n\u00eb m\u00ebnyr\u00eb t\u00eb sakt\u00eb t\u00eb gjitha sh\u00ebrbimet e rezervuara.
\nK\u00ebtu p\u00ebrfshihen jo vet\u00ebm frontet (n\u00eb t\u00eb cilat n\u00eb k\u00ebt\u00eb moment p\u00ebrdoruesit nuk hyjn\u00eb m\u00eb) dhe sh\u00ebrbimet e tyre ndihm\u00ebse (logjika e biznesit, cache etj.), por edhe bazat e ndryshme t\u00eb t\u00eb dh\u00ebnave me faktor replikimi 2 dhe m\u00eb shum\u00eb (Cassandra<\/a><\/noindex>, depozita e t\u00eb dh\u00ebnave binare<\/a><\/noindex>, depozita e ftoht\u00eb<\/a><\/noindex>, NewSQL<\/a><\/noindex> etj.).
\n15:06. M\u00eb ka ardhur informacioni se zjarri i k\u00ebrc\u00ebnon nj\u00eb nga sallat e qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave.<\/b> N\u00eb k\u00ebt\u00eb sall\u00eb nuk kemi pajisje, por fakti q\u00eb zjarri mund t\u00eb kaloj\u00eb nga \u00e7atia n\u00eb salla, ndryshon ndjesh\u00ebm situat\u00ebn.
\n(M\u00eb von\u00eb doli se nuk kishte rrezik fizik p\u00ebr sall\u00ebn, pasi ajo ishte e izoluar hermetikisht nga \u00e7atia. Rreziku ishte vet\u00ebm p\u00ebr sistemin e ftohjes s\u00eb k\u00ebsaj salle.)
\n15:07. Lejojm\u00eb ekzekutimin e komandave n\u00eb servera n\u00eb m\u00ebnyr\u00eb t\u00eb p\u00ebrshpejtuar pa kontrollime shtes\u00eb (pa kalkulatorin ton\u00eb t\u00eb dashur<\/a><\/noindex>).
\n15:08. Temperatura n\u00eb salla \u00ebsht\u00eb brenda norm\u00ebs.
\n15:12. \u00cbsht\u00eb regjistruar nj\u00eb rritje e temperatur\u00ebs n\u00eb salla.<\/b>
\n15:13. M\u00eb shum\u00eb se gjysma e serverave n\u00eb qendr\u00ebn e t\u00eb dh\u00ebnave jan\u00eb \u00e7aktivizuar. Po vazhdojm\u00eb.
\n15:16. \u00cbsht\u00eb marr\u00eb vendimi p\u00ebr t\u00eb \u00e7aktivizuar t\u00eb gjith\u00eb pajisjet.
\n15:21. Fillojm\u00eb t\u00eb \u00e7aktivizojm\u00eb energjin\u00eb n\u00eb serverat stateless pa \u00e7aktivizimin e sakt\u00eb t\u00eb aplikacionit dhe operativit.
\n15:23. P\u00ebrcaktohet nj\u00eb grup p\u00ebrgjegj\u00ebsish p\u00ebr MS SQL (jan\u00eb t\u00eb pak\u00ebt, var\u00ebsia e sh\u00ebrbimeve ndaj tyre nuk \u00ebsht\u00eb e madhe, por procedura e rikthimit t\u00eb funksionalitetit zgjat m\u00eb shum\u00eb koh\u00eb dhe \u00ebsht\u00eb m\u00eb e komplikuar se, p\u00ebr shembull, ajo e Cassandra).<\/p>\n

Depresioni<\/h4>\n

\n15:25. M\u00eb ka ardhur informacioni p\u00ebr \u00e7aktivizimin e energjis\u00eb n\u00eb kat\u00ebr salla nga 16 (Nr. 6, 7, 8, 9).<\/b> N\u00eb sallat 7 dhe 8 ndodhet pajisja jon\u00eb. Nuk ka ende informacion rreth dy sallave tona (nr. 1 dhe 3).
\nN\u00eb p\u00ebrgjith\u00ebsi, kur ndodhin zjarre, energjia elektrike nd\u00ebrpritet menj\u00ebher\u00eb, por n\u00eb k\u00ebt\u00eb rast, fal\u00eb pun\u00ebs s\u00eb koordinuar t\u00eb zjarrfik\u00ebsve dhe stafit teknik t\u00eb qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave, ajo nuk u fik n\u00eb \u00e7do vend dhe menj\u00ebher\u00eb, por sipas nevoj\u00ebs.
\n(M\u00eb von\u00eb u zbulua se energjia n\u00eb sallat 8 dhe 9 nuk u nd\u00ebrpre.)
\n15:28. Fillojm\u00eb t\u00eb rikthejm\u00eb bazat MS SQL nga backup-et n\u00eb qendrat e tjera t\u00eb t\u00eb dh\u00ebnave.
\nSa koh\u00eb do t\u00eb k\u00ebrkohet p\u00ebr k\u00ebt\u00eb? A mjafton kapaciteti i rrjetit n\u00eb t\u00ebr\u00eb itinerarin?
\n15:37. \u00cbsht\u00eb regjistruar nd\u00ebrprerja e disa pjes\u00ebve t\u00eb rrjetit.<\/b>
\nMenaxhimi dhe rrjeti i prodhimit jan\u00eb fizikisht t\u00eb izoluar nga nj\u00ebri-tjetri. N\u00ebse rrjeti i prodhimit \u00ebsht\u00eb i aksesuesh\u00ebm, at\u00ebher\u00eb mund t\u00eb hyni n\u00eb server, t\u00eb ndaloni aplikacionin dhe t\u00eb fikni sistemin operativ. N\u00ebse nuk \u00ebsht\u00eb i aksesuesh\u00ebm, at\u00ebher\u00eb mund t\u00eb hysh p\u00ebrmes IPMI, t\u00eb ndalosh aplikacionin dhe t\u00eb fik\u00ebsh sistemin operativ. N\u00ebse nuk ka asnj\u00eb nga rrjetet, nuk mund t\u00eb b\u00ebni asgj\u00eb. 'Faleminderit, kapiten!', do t\u00eb mendoni.
\n'Po ashtu, si duket, ka shum\u00eb zhurm\u00eb', ndoshta do t\u00eb mendoni.
\nE gjith\u00eb \u00e7\u00ebshtja \u00ebsht\u00eb se server\u00ebt, edhe pa zjarr, gjenerojn\u00eb nj\u00eb sasi t\u00eb madhe heat. Sakt\u00ebsisht, kur ka ftohje, ata gjenerojn\u00eb nxeht\u00ebsi, dhe kur nuk ka, ata krijojn\u00eb nj\u00eb ferr t\u00eb tmerrsh\u00ebm, i cili n\u00eb rastin m\u00eb t\u00eb mir\u00eb do t\u00eb shkrij\u00eb nj\u00eb pjes\u00eb t\u00eb pajisjeve dhe do t\u00eb fik\u00eb nj\u00eb pjes\u00eb tjet\u00ebr, por n\u00eb rastin m\u00eb t\u00eb keq\u2026 do t\u00eb shkaktoj\u00eb nj\u00eb zjarr brenda sall\u00ebs, i cili praktikisht garanton shkat\u00ebrrimin e gjith\u00e7kaje.<\/p>\n

\"A
\n
\n15:39. Regjistrojm\u00eb probleme me baz\u00ebn conf.<\/p>\n

Baza conf \u00ebsht\u00eb backend p\u00ebr sh\u00ebrbimin e nj\u00ebjt\u00eb, i cili p\u00ebrdoret nga t\u00eb gjitha aplikacionet e prodhimit p\u00ebr ndryshimin e shpejt\u00eb t\u00eb konfigurimeve. Pa k\u00ebt\u00eb baz\u00eb ne nuk mund t\u00eb menaxhojm\u00eb funksionimin e portalit, por portali mund t\u00eb funksionoj\u00eb gjithsesi.<\/p><\/blockquote>\n

\n15:41. Sensor\u00ebt e temperatur\u00ebs n\u00eb pajisjet thelb\u00ebsore t\u00eb rrjetit regjistrojn\u00eb lexime t\u00eb af\u00ebrta me ata t\u00eb lejuar. Ky \u00ebsht\u00eb nj\u00eb kuti q\u00eb z\u00eb nj\u00eb raft dhe siguron funksionimin e t\u00eb gjith\u00eb rrjeteve brenda qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave.<\/p>\n

\"A
\n
\n15:42. Issue tracker dhe wiki nuk jan\u00eb t\u00eb aksesueshme, kalojm\u00eb n\u00eb standby.
\nKjo nuk \u00ebsht\u00eb prodhim, por n\u00eb rast emergjence aksesueshm\u00ebria e \u00e7do baze njohurish mund t\u00eb jet\u00eb kritike.
\n15:50. Nj\u00eb nga sistemet e monitorimit u ndal.
\nAta jan\u00eb disa, dhe ata mbulojn\u00eb aspekte t\u00eb ndryshme t\u00eb funksionimit t\u00eb sh\u00ebrbimeve. Disa prej tyre jan\u00eb t\u00eb configuruar p\u00ebr t\u00eb punuar autonomisht brenda \u00e7do qendre t\u00eb t\u00eb dh\u00ebnave (do t\u00eb thot\u00eb se monitorojn\u00eb vet\u00ebm qendr\u00ebn e tyre t\u00eb t\u00eb dh\u00ebnave), nd\u00ebrsa t\u00eb tjer\u00ebt p\u00ebrb\u00ebhen nga komponente t\u00eb shp\u00ebrndara, q\u00eb p\u00ebsojn\u00eb transparenc\u00eb n\u00eb rast humbjeje t\u00eb ndonj\u00eb qendre t\u00eb t\u00eb dh\u00ebnave.
\nN\u00eb k\u00ebt\u00eb rast, ka pushuar s\u00eb funksionuari sistemi i zbulimit t\u00eb anomaliave t\u00eb treguesve t\u00eb logjik\u00ebs s\u00eb biznesit<\/a><\/noindex>, i cili operon n\u00eb modin master-standby. Kaluam n\u00eb standby.<\/p>\n

Pranimi<\/h4>\n

\n15:51. P\u00ebrdor\u00ebm IPMI p\u00ebr t\u00eb fikur t\u00eb gjith\u00eb serverat pa p\u00ebrfundim t\u00eb duhur, p\u00ebrve\u00e7 MS SQL.
\nA jeni t\u00eb gatsh\u00ebm p\u00ebr menaxhimin masiv t\u00eb server\u00ebve p\u00ebrmes IPMI n\u00eb raste nevoje?<\/p>\n

Ky \u00ebsht\u00eb momenti ku shp\u00ebtimi i pajisjeve n\u00eb qendr\u00ebn e t\u00eb dh\u00ebnave n\u00eb k\u00ebt\u00eb faz\u00eb \u00ebsht\u00eb p\u00ebrfunduar. \u00c7far\u00eb mund t\u00eb b\u00ebhej \u00ebsht\u00eb b\u00ebr\u00eb. Disa koleg\u00eb mund t\u00eb pushojn\u00eb.<\/i>
\n16:13. Kemi marr\u00eb informacion se tubat e freonit nga kondicioner\u00ebt n\u00eb \u00e7ati jan\u00eb thyer - kjo do ta vonoj\u00eb hapjen e qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave pas eliminimit t\u00eb zjarrit.<\/b>
\n16:19. Sipas t\u00eb dh\u00ebnave t\u00eb marra nga stafi teknik i qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave, ndalimi i rritjes s\u00eb temperatur\u00ebs n\u00eb sallat ka ndodhur.
\n17:10. Rind\u00ebrtuam funksionimin e baz\u00ebs conf. Tani mund t\u00eb ndryshojm\u00eb konfigurimet e aplikacioneve.
\nPse \u00ebsht\u00eb kaq e r\u00ebnd\u00ebsishme, n\u00ebse gjith\u00e7ka \u00ebsht\u00eb e vijueshme dhe punon edhe pa nj\u00eb qend\u00ebr t\u00eb dh\u00ebnash?
\nS\u00eb pari, jo gjith\u00e7ka \u00ebsht\u00eb e vijueshme. Ka sh\u00ebrbime t\u00eb ndryshme dyt\u00ebsore q\u00eb p\u00ebr momentin nuk p\u00ebrballojn\u00eb mjaft mir\u00eb d\u00ebshtimin e qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave, dhe ka baza n\u00eb modin master-standby. Mund\u00ebsia p\u00ebr t\u00eb menaxhuar konfigurimet lejon q\u00eb t\u00eb b\u00ebhet \u00e7do gj\u00eb e nevojshme p\u00ebr t\u00eb minimizuar ndikimin e pasojave t\u00eb aksidenteve n\u00eb p\u00ebrdorues edhe n\u00eb kushte t\u00eb v\u00ebshtira.
\nS\u00eb dyti, b\u00ebhet e qart\u00eb se n\u00eb or\u00ebt e ardhshme puna e qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave nuk do t\u00eb rikthehet plot\u00ebsisht, ndaj ishte e nevojshme t\u00eb merreshin masa q\u00eb mosshp\u00ebrndarja afatgjate e kopjeve t\u00eb sillte probleme t\u00eb tjera si mbushja e disqeve n\u00eb qendrat e mbetura t\u00eb t\u00eb dh\u00ebnave.
\n17:29. \u00cbsht\u00eb koha p\u00ebr pic\u00eb! Ne kemi njer\u00ebz q\u00eb punojn\u00eb, jo robot\u00eb.<\/p>\n

\"A
\n<\/p>\n

Rehabilitimi<\/h4>\n

\n18:02. N\u00eb sallat \u2116\u21168 (tona), 9, 10 dhe 11 temperatura u stabilizua. N\u00eb nj\u00ebr\u00ebn nga sallat q\u00eb mbeten t\u00eb fikura (\u21167), ndodhet pajisja jon\u00eb, dhe temperatura atje vazhdon t\u00eb rritet.
\n18:31. Dh\u00ebn\u00eb lejen p\u00ebr t\u00eb ndezur pajisjet n\u00eb sallat \u2116\u21161 dhe 3 - k\u00ebto sallat nuk u prek\u00ebn nga zjarri.<\/p>\n

Aktualisht po b\u00ebhet nisja e server\u00ebve n\u00eb sallat \u2116\u21161, 3, 8, duke filluar nga m\u00eb t\u00eb kritikuarit. Po kontrollohet sakt\u00ebsia e funksionimit t\u00eb t\u00eb gjitha sh\u00ebrbimeve t\u00eb nisura. Ende ka probleme me sall\u00ebn \u21167.<\/i><\/p>\n

18:44. Personeli teknik i qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave zbuloi se n\u00eb sall\u00ebn \u21167 (ku \u00ebsht\u00eb vet\u00ebm pajisja jon\u00eb) shum\u00eb server\u00eb nuk jan\u00eb fikur. Sipas t\u00eb dh\u00ebnave tona, aty mbeten t\u00eb ndezur 26 server\u00eb. Pas nj\u00eb kontrolli t\u00eb p\u00ebrs\u00ebritur zbulojm\u00eb 58 server\u00eb.
\n20:18. Personeli teknik i qendr\u00ebs s\u00eb t\u00eb dh\u00ebnave po ajros sall\u00ebn pa kondicioner\u00eb p\u00ebrmes kanaleve mobile t\u00eb ajrit, t\u00eb vendosura p\u00ebrmes korridoreve.
\n23:08. L\u00ebshuam adminin e par\u00eb p\u00ebr n\u00eb sht\u00ebpi. Disa duhet t\u00eb flen\u00eb gjat\u00eb nat\u00ebs, q\u00eb nes\u00ebr t\u00eb vazhdojm\u00eb pun\u00ebt. M\u00eb pas, l\u00ebshojm\u00eb edhe nj\u00eb pjes\u00eb t\u00eb admin\u00ebve dhe zhvilluesve.
\n02:56. Aktivizuam gjith\u00e7ka q\u00eb mund t\u00eb aktivizohej. Po b\u00ebjm\u00eb nj\u00eb kontroll t\u00eb madh t\u00eb t\u00eb gjitha sh\u00ebrbimeve me teste t\u00eb automatizuara.<\/p>\n

\"A
\n
\n03:02. Kondicionimi n\u00eb sall\u00ebn e fundit, \u21167, \u00ebsht\u00eb rikthyer.
\n03:36. Vendos\u00ebm frontet n\u00eb qendr\u00ebn e t\u00eb dh\u00ebnave n\u00eb rotacion n\u00eb DNS. Nga ky moment fillon t\u00eb arrij\u00eb trafiku i p\u00ebrdoruesve.
\nPo l\u00ebshojm\u00eb pjes\u00ebn m\u00eb t\u00eb madhe t\u00eb ekipit t\u00eb administrator\u00ebve p\u00ebr n\u00eb sht\u00ebpi. Por disa persona i l\u00ebm\u00eb.<\/p>\n

Nj\u00eb FAQ i vog\u00ebl:
\nQ: \u00c7far\u00eb ndodhi nga 18:31 deri n\u00eb 02:56?
\nA: Duke ndjekur \"Planin e Veprimit n\u00eb Rast t\u00eb Aksidentit\", ne aktivizojm\u00eb t\u00eb gjitha sh\u00ebrbimet, duke filluar nga m\u00eb t\u00eb r\u00ebnd\u00ebsishmet. N\u00eb k\u00ebt\u00eb rast, koordinatori n\u00eb bised\u00eb i jep sh\u00ebrbimin administratorit t\u00eb lir\u00eb, i cili kontrollon n\u00ebse jan\u00eb nisur OS dhe aplikacioni, a ka ndonj\u00eb gabim, a jan\u00eb treguesit n\u00eb rregull. Pas p\u00ebrfundimit t\u00eb aktivizimit, ai njofton n\u00eb bised\u00eb se \u00ebsht\u00eb i lir\u00eb dhe merr nga koordinatori nj\u00eb sh\u00ebrbim t\u00eb ri.
\nProcesi ngadal\u00ebsohet gjithashtu nga hardueri i d\u00ebshtuar. Edhe n\u00ebse ndalimi i OS dhe fikja e server\u00ebve kalojn\u00eb si\u00e7 duhet, disa server\u00eb nuk kthehen p\u00ebr shkak t\u00eb disqeve, memories, shasis\u00eb q\u00eb d\u00ebshtojn\u00eb papritur. N\u00eb rast t\u00eb humbjes s\u00eb energjis\u00eb elektrike, p\u00ebrqindja e d\u00ebshtimeve rritet.
\nQ: Pse nuk mund ta aktivizoni gjith\u00e7ka nj\u00ebher\u00ebsh dhe pastaj t\u00eb riparoni at\u00eb q\u00eb del n\u00eb monitorim?
\nA: \u00c7do gj\u00eb duhet b\u00ebr\u00eb gradualisht, sepse ka var\u00ebsi midis sh\u00ebrbimeve. Dhe duhet kontrolluar gjith\u00e7ka menj\u00ebher\u00eb, pa pritur monitorimin \u2014 sepse \u00ebsht\u00eb m\u00eb mir\u00eb t\u00eb zgjidhen problemet menj\u00ebher\u00eb, pa pritur q\u00eb ato t\u00eb p\u00ebrkeq\u00ebsohen.<\/p><\/blockquote>\n

\n7:40. Administratori i fundit (koordinatori) u shkua p\u00ebr t\u00eb fjetur. Pun\u00ebt e dit\u00ebs s\u00eb par\u00eb p\u00ebrfundojn\u00eb.
\n8:09. Inxhinier\u00ebt e par\u00eb dhe administrator\u00ebt n\u00eb qendrat e t\u00eb dh\u00ebnave (p\u00ebrfshir\u00eb koordinatorin e ri) filluan pun\u00ebn p\u00ebr rikuperim.
\n09:37. Filluam ng\u00ebr\u00e7in e sall\u00ebs \u21167 (salla e fundit).
\nParalelisht vazhdojm\u00eb t\u00eb rikuperojm\u00eb at\u00eb q\u00eb nuk e p\u00ebrfunduam n\u00eb sallat e tjera: z\u00ebvend\u00ebsimin e disqeve\/memories\/serveryve, riparimin e gjith\u00e7kaje q\u00eb \"digjet\" n\u00eb monitorim, rikthimin e roleve n\u00eb skemat master-standby dhe ndihma me gj\u00ebra t\u00eb tjera, q\u00eb megjithat\u00eb jan\u00eb mjaft.
\n17:08. Lejojm\u00eb t\u00eb gjitha punimet standarde me prodhimin.
\n21:45. Pun\u00ebt e dit\u00ebs s\u00eb dyt\u00eb p\u00ebrfunduan.
\n09:45. Sot \u00ebsht\u00eb e premte. N\u00eb monitorim ende ka shum\u00eb probleme t\u00eb vogla. T\u00eb shtun\u00ebn po afrohet, t\u00eb gjith\u00eb duan t\u00eb pushojn\u00eb. Vazhdoni t\u00eb riparoni masivisht gjith\u00e7ka q\u00eb \u00ebsht\u00eb e mundur. Detyrat standarde t\u00eb administrator\u00ebve q\u00eb mund t\u00eb ishin shtyr\u00eb jan\u00eb shtyr\u00eb. Koordinatori \u00ebsht\u00eb i ri.
\n15:40. Papritmas, gjysma e stackut t\u00eb pajisjeve rrjetore n\u00eb nj\u00eb qend\u00ebr tjet\u00ebr t\u00eb t\u00eb dh\u00ebnave u ristartua. Nxorr\u00ebm p\u00ebrjashtim nga rotacioni p\u00ebr minimizimin e riskut. Nuk kishte ndonj\u00eb efekt p\u00ebr p\u00ebrdoruesit. M\u00eb von\u00eb u zbulua se kjo ishte nj\u00eb \u00e7\u00ebshtje e shasis\u00eb. Koordinatori po punon me riparimin e dy emergjencave.
\n17:17. Funksionimi i rrjetit n\u00eb qendr\u00ebn tjet\u00ebr t\u00eb t\u00eb dh\u00ebnave \u00ebsht\u00eb rikthyer, gjith\u00e7ka \u00ebsht\u00eb kontrolluar. Qendra e t\u00eb dh\u00ebnave \u00ebsht\u00eb futur n\u00eb rotacion.
\n18:29. Pun\u00ebt e dit\u00ebs s\u00eb tret\u00eb dhe n\u00eb p\u00ebrgjith\u00ebsi rikuperimi pas emergjenc\u00ebs jan\u00eb p\u00ebrfunduar.<\/p>\n

Pasth\u00ebnie<\/h4>\n

\n04.04.2013. n\u00eb dit\u00ebn e gabimit 404<\/a><\/noindex>, \"Odnoklassniki\" p\u00ebrjetoi emergjenc\u00ebn m\u00eb t\u00eb madhe<\/a><\/noindex> \u2014 p\u00ebr tre dit\u00eb portali ishte totalisht apo pjes\u00ebrisht i pap\u00ebrshtatsh\u00ebm. Gjat\u00eb gjith\u00eb k\u00ebsaj kohe, m\u00eb shum\u00eb se 100 persona nga qytete t\u00eb ndryshme, nga kompani t\u00eb ndryshme (edhe nj\u00ebher\u00eb faleminderit shum\u00eb!), riparuan mij\u00ebra servera, n\u00eb distanc\u00eb dhe drejtp\u00ebrdrejt n\u00eb qendrat e t\u00eb dh\u00ebnave, n\u00eb m\u00ebnyr\u00eb manuale dhe automatike.
\nNe kemi nxjerr\u00eb p\u00ebrfundime. Q\u00eb t\u00eb mos ndodhi p\u00ebrs\u00ebri di\u00e7ka e till\u00eb, ne kemi realizuar dhe vazhdojm\u00eb t\u00eb realizojm\u00eb deri m\u00eb sot pun\u00eb t\u00eb gjera.<\/p>\n

Cilat jan\u00eb dallimet kryesore midis k\u00ebsaj emergjence dhe asaj 404?<\/p>\n