{"id":37745,"date":"2019-10-31T22:19:28","date_gmt":"2019-10-31T19:19:28","guid":{"rendered":"https:\/\/prohoster.info\/blog\/balansirovka-nagruzki-v-openstack-chast-2\/"},"modified":"2019-10-31T22:19:28","modified_gmt":"2019-10-31T19:19:28","slug":"balansirovka-nagruzki-v-openstack-chast-2","status":"publish","type":"post","link":"https:\/\/prohoster.info\/ro\/blog\/administrirovanie\/balansirovka-nagruzki-v-openstack-chast-2","title":{"rendered":"\u00cenc\u0103rcarea echilibrat\u0103 \u00een Openstack (Partea 2)","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>\u00cen <noindex><a rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/rostelecom\/blog\/461483\/\">articolul precedent<\/a><\/noindex> Am discutat despre \u00eencerc\u0103rile de a utiliza Watcher \u0219i am prezentat un raport al testelor. Astfel de teste le efectu\u0103m periodic pentru echilibrare \u0219i alte func\u021bii critice ale unui cloud corporativ sau de operator de mari dimensiuni.<\/p>\n<p>Gradul ridicat de complexitate al problemei poate necesita mai multe articole pentru a descrie proiectul nostru. Ast\u0103zi public\u0103m al doilea articol din ciclu, dedicat echilibr\u0103rii ma\u0219inilor virtuale \u00een cloud. <\/p>\n<h3>Pu\u021bin\u0103 terminologie<\/h3>\n<p>\nCompania VmWare a introdus utilitarul DRS (Distributed Resource Scheduler) pentru a echilibra sarcinile de lucru \u00een mediul de virtualizare dezvoltat \u0219i oferit de aceasta. <\/p>\n<p>A\u0219a cum scrie <noindex><a rel=\"nofollow\" href=\"https:\/\/searchvmware.techtarget.com\/definition\/VMware-DRS\">searchvmware.techtarget.com\/definition\/VMware-DRS<\/a><\/noindex><br \/>\n<i>\u00abVMware DRS (Planificatorul de resurse distribuite) este un utilitar care echilibreaz\u0103 sarcinile de calcul cu resursele disponibile \u00een mediul virtual. Utilitarul face parte din pachetul de virtualizare numit VMware Infrastructure.<\/p>\n<p>Cu ajutorul VMware DRS, utilizatorii definesc reguli pentru distribuirea resurselor fizice \u00eentre ma\u0219inile virtuale (MV). Utilitarul poate fi configurat pentru gestionarea manual\u0103 sau automat\u0103. Puzzile de resurse VMware pot fi u\u0219or ad\u0103ugate, eliminate sau reorganizate. La cerere, puzzile de resurse pot fi izolate \u00eentre diferitele unit\u0103\u021bi de afaceri. Dac\u0103 sarcina de lucru pe una sau mai multe ma\u0219ini virtuale se schimb\u0103 brusc, VMware DRS redistribuie ma\u0219inile virtuale \u00eentre serverele fizice. Dac\u0103 sarcina de lucru total\u0103 scade, unele servere fizice pot fi temporar oprite, iar sarcina de lucru consolidat\u0103.\u00bb<\/i><\/p>\n<h3>De ce este necesar\u0103 echilibrarea?<\/h3>\n<p>\n<noindex><a rel=\"nofollow\" name=\"habracut\"><\/a><\/noindex><br \/>\n\u00cen opinia noastr\u0103, DRS este o func\u021bie esen\u021bial\u0103 a cloud-ului, de\u0219i aceasta nu \u00eenseamn\u0103 c\u0103 DRS trebuie utilizat \u00eentotdeauna \u0219i oriunde. \u00cen func\u021bie de scopul \u0219i nevoile cloud-ului, pot exista cerin\u021be diferite pentru DRS \u0219i metodele de echilibrare. Exist\u0103 probabil situa\u021bii c\u00e2nd echilibrarea nu este deloc necesar\u0103. Sau chiar d\u0103un\u0103toare.<\/p>\n<p>Pentru a \u00een\u021belege mai bine unde \u0219i pentru ce clien\u021bi este necesar DRS, s\u0103 lu\u0103m \u00een considerare obiectivele \u0219i sarcinile lor. Cloud-urile pot fi \u00eemp\u0103r\u021bite \u00een publice \u0219i private. Iat\u0103 principalele diferen\u021be \u00eentre aceste cloud-uri \u0219i obiectivele clien\u021bilor.<\/p>\n<p>Cloud-uri private \/ Clien\u021bi corporativi mari<br \/>\nCloud-uri publice \/ \u00centreprinderi mici \u0219i mijlocii, persoane<\/p>\n<p>Criteriul principal \u0219i obiectivele operatorului<br \/>\nFurnizarea unui serviciu sau produs de \u00eencredere<br \/>\nReducerea costurilor serviciilor \u00een lupta pe pia\u021ba competitiv\u0103<\/p>\n<p>Cerin\u021bele pentru serviciu<br \/>\nFiabilitate la toate nivelurile \u0219i \u00een toate componentele sistemului<\/p>\n<p>Performan\u021b\u0103 garantat\u0103<\/p>\n<p>Prioritizarea ma\u0219inilor virtuale \u00een mai multe categorii\u00a0<\/p>\n<p>Securitatea informa\u021bional\u0103 \u0219i fizic\u0103 a datelor<\/p>\n<p>SLA \u0219i suport 24\/7<br \/>\nMaxima simplificare \u00een ob\u021binerea serviciului<\/p>\n<p>Servicii relativ simple<\/p>\n<p>Responsabilitatea pentru date revine clientului<\/p>\n<p>Prioritizarea VM nu este necesar\u0103<\/p>\n<p>Securitatea informa\u021bional\u0103 la nivelul serviciilor standard, responsabilitatea revine clientului<\/p>\n<p>Pot exista defec\u021biuni<\/p>\n<p>F\u0103r\u0103 SLA, calitatea nu este garantat\u0103<\/p>\n<p>Suport prin e-mail<\/p>\n<p>Backup-ul nu este obligatoriu<\/p>\n<p>Particularit\u0103\u021bile clientului<br \/>\nO gam\u0103 foarte larg\u0103 de aplica\u021bii.<\/p>\n<p>Aplica\u021bii legate de companie.<\/p>\n<p>Arhitecturi personalizate complexe pentru fiecare client.<\/p>\n<p>Reguli de afinitate.<\/p>\n<p>Func\u021bionarea software-ului f\u0103r\u0103 oprire \u00een modul 7x24.\u00a0<\/p>\n<p>Instrumente de backup \u201e\u00een timpul func\u021bion\u0103rii\u201d.<\/p>\n<p>\u00cenc\u0103rc\u0103tur\u0103 ciclic\u0103 predictibil\u0103 a clientului.<br \/>\nAplica\u021bii standard \u2013 echilibrarea re\u021belei, Apache, WEB, VPN, SQL<\/p>\n<p>Posibilitatea de a opri aplica\u021bia pentru o perioad\u0103 de timp<\/p>\n<p>Se permite distribuirea aleatorie a VM \u00een cloud<\/p>\n<p>Backup efectuat de client<\/p>\n<p>\u00cenc\u0103rc\u0103tur\u0103 medie previzibil\u0103 statistic pentru un num\u0103r mare de clien\u021bi.<\/p>\n<p>Consecin\u021be pentru arhitectur\u0103<br \/>\nGeoclusterizare<\/p>\n<p>Stocare centralizat\u0103 sau distribuit\u0103 S\text{H\u00ce}<\/p>\n<p>R\text{SRK} rezervabil\u0103<br \/>\nStocarea local\u0103 a datelor pe nodurile de calcul<\/p>\n<p>Obiectivele echilibr\u0103rii<br \/>\nDistribu\u021bia uniform\u0103 a sarcinii<\/p>\n<p>Maxima reactivitate a aplica\u021biilor\u00a0<\/p>\n<p>Timp minim de \u00eent\u00e2rziere la echilibrare<\/p>\n<p>Echilibrarea doar \u00een caz de necesitate evident\u0103<\/p>\n<p>Dezvoltarea unei p\u0103r\u021bi a echipamentului pentru \u00eentre\u021binere preventiv\u0103<br \/>\nReducerea costului serviciului \u0219i a cheltuielilor operatorului\u00a0<\/p>\n<p>Dezactivarea unor resurse \u00een cazul unei \u00eenc\u0103rc\u0103turi reduse<\/p>\n<p>Economisirea energiei electrice<\/p>\n<p>Reducerea costurilor pentru personal<\/p>\n<h3> Facem urm\u0103toarele concluzii pentru noi:<\/h3>\n<p>\n<b>Pentru cloud-uri private<\/b>, ofere\u021bi institu\u021biilor mari, DRS poate fi aplicat \u021bin\u00e2nd cont de restric\u021bii:<\/p>\n<ul>\n<li>securitatea informa\u021bional\u0103 \u0219i respectarea regulilor de afinitate la echilibrare;<\/li>\n<li>disponibilitatea \u00eentr-un rezervor a unui volum suficient de resurse \u00een caz de accident;<\/li>\n<li>datele ma\u0219inilor virtuale se afl\u0103 pe un S\text{H\u00ce} centralizat sau distribuit;<\/li>\n<li>distribuirea \u00een timp a procedurilor de administrare, backup \u0219i echilibrare;<\/li>\n<li>echilibrarea doar \u00een cadrul agregatului de gazde ale clientului;<\/li>\n<li>echilibrarea doar \u00een caz de un dezechilibru semnificativ, cele mai eficiente \u0219i sigure migra\u021bii VM (deoarece migrarea poate fi \u0219i nereu\u0219it\u0103);<\/li>\n<li>echilibrarea \u00een raport cu ma\u0219inile virtuale \u00abcalme\u00bb (migrarea ma\u0219inilor virtuale \u00abzgomotoase\u00bb poate dura foarte mult);<\/li>\n<li>echilibrarea av\u00e2nd \u00een vedere \u00abcostul\u00bb \u2014 \u00eenc\u0103rcarea pe sistemul de stocare \u0219i re\u021bea (\u00een arhitecturi personalizate pentru clien\u021bi mari);<\/li>\n<li>echilibrarea cu considera\u021bie pentru particularit\u0103\u021bile comportamentului fiec\u0103rei VM;<\/li>\n<li>echilibrarea preferabil \u00een afara orelor de lucru (noaptea, \u00een weekend, s\u0103rb\u0103tori).<\/li>\n<\/ul>\n<p>\n<b>Pentru cloud-urile publice<\/b>, care ofer\u0103 servicii clien\u021bilor mici, DRS poate fi aplicat mult mai frecvent, cu func\u021bionalit\u0103\u021bi extinse:<\/p>\n<ul>\n<li>lipsa restric\u021biilor de securitate informa\u021bional\u0103 \u0219i a regulilor de afinitate; <\/li>\n<li>echilibrarea \u00een cadrul cloud-ului;<\/li>\n<li>echilibrarea \u00een orice moment rezonabil;<\/li>\n<li>echilibrarea oric\u0103rei VM; <\/li>\n<li>echilibrarea ma\u0219inilor virtuale \u00abzgomotoase\u00bb (pentru a nu deranja restul); <\/li>\n<li>datele ma\u0219inilor virtuale se afl\u0103 adesea pe discuri locale;<\/li>\n<li>considerarea performan\u021bei medii a sistemului de stocare \u0219i a re\u021belei (arhitectura cloud-ului este unitar\u0103);<\/li>\n<li>echilibrarea conform regulilor generalizate \u0219i statisticii comportamentului DC.<\/li>\n<\/ul>\n<p><\/p>\n<h2>Complexitatea problemei<\/h2>\n<p>\nComplexitatea echilibr\u0103rii const\u0103 \u00een faptul c\u0103 DRS trebuie s\u0103 func\u021bioneze cu un num\u0103r mare de factori necunoscu\u021bi:<\/p>\n<ul>\n<li>comportamentul utilizatorilor fiec\u0103rui sistem informa\u021bional al clien\u021bilor;<\/li>\n<li>algoritmii de func\u021bionare ai serverelor sistemelor informa\u021bionale;<\/li>\n<li>comportamentul serverelor DBMS;<\/li>\n<li>\u00eenc\u0103rcarea pe resursele de calcul, sistemul de stocare, re\u021bea; <\/li>\n<li>interac\u021biunea serverelor \u00eentre ele \u00een competi\u021bia pentru resursele cloud-ului.<\/li>\n<\/ul>\n<p>\n\u00cenc\u0103rcarea unui num\u0103r mare de servere virtuale pentru aplica\u021bii \u0219i baze de date pe resursele cloud-ului se desf\u0103\u0219oar\u0103 \u00een timp, consecin\u021bele put\u00e2nd ap\u0103rea \u0219i suprapune unele peste altele cu efecte imprevizibile la momente imprevizibile. Chiar \u0219i pentru gestionarea proceselor relativ simple (de exemplu, gestionarea motorului, sistemului de \u00eenc\u0103lzire pe baz\u0103 de ap\u0103 al casei), sistemele de reglare automat\u0103 trebuie s\u0103 foloseasc\u0103 algoritmi complica\u021bi <noindex><a rel=\"nofollow\" href=\"https:\/\/ru.wikipedia.org\/wiki\/%D0%9F%D0%98%D0%94-%D1%80%D0%B5%D0%B3%D1%83%D0%BB%D1%8F%D1%82%D0%BE%D1%80\">propor\u021bional-integrali-differentiating <\/a><\/noindex>cu feedback. <\/p>\n<p><img decoding=\"async\" alt=\"\u00cenc\u0103rcarea echilibrat\u0103 \u00een Openstack (Partea 2)\" src=\"\/wp-content\/uploads\/2019\/09\/5270ddb943db9b4921a887e67c4f2a8f.png\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nSarcina noastr\u0103 este mult mai complex\u0103, iar exist\u0103 riscul ca sistemul s\u0103 nu poat\u0103 realiza echilibrarea \u00eenc\u0103rc\u0103turii \u00een valori stabilite \u00eentr-un timp rezonabil, chiar dac\u0103 nu apar interven\u021bii externe din partea utilizatorilor. <\/p>\n<p><img decoding=\"async\" alt=\"\u00cenc\u0103rcarea echilibrat\u0103 \u00een Openstack (Partea 2)\" src=\"\/wp-content\/uploads\/2019\/09\/5214b3af233e7c718fd40e6336472903.png\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<\/p>\n<h2>Istoria dezvolt\u0103rilor noastre<\/h2>\n<p>\nPentru a rezolva aceast\u0103 problem\u0103, am decis s\u0103 nu \u00eencepem de la zero, ci s\u0103 ne baz\u0103m pe experien\u021ba existent\u0103, colabor\u00e2nd cu speciali\u0219ti cu experien\u021b\u0103 \u00een acest domeniu. Din fericire, \u00een\u021belegerea problemelor noastre a fost complet aliniat\u0103. <\/p>\n<h3>Etapa 1<\/h3>\n<p>\nAm folosit un sistem bazat pe tehnologia re\u021belelor neuronale \u0219i am \u00eencercat s\u0103 ne optimiz\u0103m resursele pe baza acestuia.<\/p>\n<p>Interesul acestei etape a consistat \u00een testarea unei noi tehnologii, iar importan\u021ba sa a fost aplicarea unei abord\u0103ri neobi\u0219nuite pentru rezolvarea problemei, unde, \u00een condi\u021bii egale, abord\u0103rile standard s-au epuizat practic.<\/p>\n<p>Am lansat sistemul \u0219i, \u00eentr-adev\u0103r, a \u00eenceput echilibrarea. Scala cloud-ului nostru nu ne-a permis s\u0103 ob\u021binem rezultate optimiste, a\u0219a cum au declarat dezvoltatorii, dar a fost clar c\u0103 echilibrarea func\u021bioneaz\u0103. <\/p>\n<p>\u00cen acela\u0219i timp, aveam restric\u021bii destul de serioase:<\/p>\n<ul>\n<li>Pentru a antrena re\u021beaua neuronal\u0103, este necesar ca ma\u0219inile virtuale s\u0103 func\u021bioneze f\u0103r\u0103 modific\u0103ri semnificative timp de s\u0103pt\u0103m\u00e2ni sau luni.<\/li>\n<li>Algoritmul este conceput pentru optimizare pe baza analizei datelor 'istorice' anterioare.<\/li>\n<li>Pentru a antrena re\u021beaua neuronal\u0103 este nevoie de un volum suficient de mare de date \u0219i resurse de calcul. <\/li>\n<li>Optimizarea \u0219i echilibrarea pot fi realizate relativ rar \u2013 la c\u00e2teva ore, ceea ce este evident insuficient. <\/li>\n<\/ul>\n<p><\/p>\n<h3>Etapa 2<\/h3>\n<p>\nDat fiind c\u0103 situa\u021bia nu ne mul\u021bumea, am decis s\u0103 modific\u0103m sistemul \u0219i, pentru aceasta, trebuie s\u0103 r\u0103spundem la <b>\u00eentrebarea principal\u0103<\/b> \u2013 pentru cine o facem?<\/p>\n<p>Mai \u00eent\u00e2i \u2013 pentru clien\u021bii corporativi. \u00censeamn\u0103 c\u0103 avem nevoie de un sistem care s\u0103 func\u021bioneze rapid, cu acele restric\u021bii corporative care simplific\u0103 doar implementarea.<\/p>\n<p><b>A doua \u00eentrebare<\/b> \u2013 ce \u00een\u021belegem prin cuv\u00e2ntul \u201erapid\u201d? \u00cen urma unor dezbateri scurte, am decis c\u0103 putem s\u0103 ne referim la un timp de reac\u021bie de 5 \u2013 10 minute, astfel \u00eenc\u00e2t fluctua\u021biile pe termen scurt s\u0103 nu introduc\u0103 sistemul \u00een rezonan\u021b\u0103.<\/p>\n<p><b>A treia \u00eentrebare<\/b> \u2013 ce dimensiune a cantit\u0103\u021bii de servere echilibrate s\u0103 alegem?<br \/>\nAceast\u0103 problem\u0103 s-a rezolvat de la sine. De obicei, clien\u021bii nu fac grupuri de servere foarte mari, iar acest lucru este conform recomand\u0103rilor din articol de a limita grupurile la 30-40 de servere.<\/p>\n<p>\u00cen plus, prin segmentarea grupului de servere, simplific\u0103m sarcina algoritmului de balansare.<\/p>\n<p><b>\u00centrebarea patra<\/b> \u2013 c\u00e2t de mult ne convine re\u021beaua neuronal\u0103 cu procesul ei lung de \u00eenv\u0103\u021bare \u0219i balans\u0103rile rare? Am decis s\u0103 renun\u021b\u0103m la ea \u00een favoarea algoritmilor opera\u021bionali mai simpli, pentru a ob\u021bine rezultate \u00een c\u00e2teva secunde. <\/p>\n<p><img decoding=\"async\" alt=\"\u00cenc\u0103rcarea echilibrat\u0103 \u00een Openstack (Partea 2)\" src=\"\/wp-content\/uploads\/2019\/09\/73fb927b304ca5d61090a5c045568931.png\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nO descriere a sistemului care folose\u0219te astfel de algoritmi \u0219i dezavantajele acestuia poate fi consultat\u0103 <noindex><a rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/post\/416391\/\">aici <\/a><\/noindex><\/p>\n<p>Am implementat \u0219i lansat acest sistem \u0219i am ob\u021binut rezultate promi\u021b\u0103toare \u2013 \u00een prezent, acesta analizeaz\u0103 \u00een mod regulat sarcina din cloud \u0219i ofer\u0103 recomand\u0103ri pentru mutarea ma\u0219inilor virtuale, care sunt \u00een mare parte corecte. Chiar \u0219i acum este evident c\u0103 putem ob\u021bine 10-15% eliberare de resurse pentru noi ma\u0219ini virtuale, \u00eembun\u0103t\u0103\u021bind calitatea func\u021bion\u0103rii celor existente.<\/p>\n<p><noindex><a rel=\"nofollow\" href=\"https:\/\/habrastorage.org\/webt\/ie\/ky\/wa\/iekywainupnidp_73j55hc1dbnw.png\"><img decoding=\"async\" alt=\"\u00cenc\u0103rcarea echilibrat\u0103 \u00een Openstack (Partea 2)\" src=\"\/wp-content\/uploads\/2019\/09\/f1f0919ae653dc9b8c78b4d32d839313.png\" style=\"display:block;margin: 0 auto;\" \/><\/a><\/noindex><\/p>\n<p>Atunci c\u00e2nd se detecteaz\u0103 un dezechilibru la nivel de RAM sau CPU, sistemul d\u0103 comenzi planificatorului Tionix pentru a efectua migrarea \u00een timpul func\u021bion\u0103rii a ma\u0219inilor virtuale necesare. A\u0219a cum se poate observa din sistemul de monitorizare, ma\u0219ina virtual\u0103 s-a mutat de la un host (superior) la altul (inferior) \u0219i a eliberat memorie pe hostul superior (evidentiat cu cercuri galbene), ocup\u00e2nd-o corespunz\u0103tor pe cel inferior (evidentiat cu cercuri albe).<\/p>\n<p>Acum ne str\u0103duim s\u0103 evalu\u0103m mai exact eficien\u021ba algoritmului \u00een func\u021biune \u0219i \u00eencerc\u0103m s\u0103 identific\u0103m eventualele sale erori.<\/p>\n<h3>Etapa 3<\/h3>\n<p>\nAr p\u0103rea c\u0103 aici putem s\u0103 ne lini\u0219tim, a\u0219tept\u00e2nd eficien\u021ba dovedit\u0103 \u0219i s\u0103 \u00eencheiem subiectul.<br \/>\nDar suntem \u00eempin\u0219i spre un nou stadiu de urm\u0103toarele oportunit\u0103\u021bi evidente de optimizare<\/p>\n<ol>\n<li>Statisticile, de exemplu, <noindex><a rel=\"nofollow\" href=\"https:\/\/www.cpubenchmark.net\/multi_cpu.html\">aici <\/a><\/noindex>\u0219i <noindex><a rel=\"nofollow\" href=\"https:\/\/www.cpubenchmark.net\/high_end_cpus.html\">aici <\/a><\/noindex>arat\u0103 c\u0103 sistemele cu dou\u0103 \u0219i patru procesoare sunt semnificativ mai pu\u021bin performante dec\u00e2t cele cu un procesor. Asta \u00eenseamn\u0103 c\u0103 to\u021bi utilizatorii ob\u021bin de la sistemele multi-procesoare CPU, RAM, SSD, LAN, FC o randament considerabil mai mic \u00een compara\u021bie cu sistemele uniprocesor. <\/li>\n<li>\u00cens\u0103 planificatoarele de resurse pot func\u021biona cu erori grave, <noindex><a rel=\"nofollow\" href=\"https:\/\/www.ece.ubc.ca\/~sasha\/papers\/eurosys16-final29.pdf\">iat\u0103 unul dintre articole<\/a><\/noindex> pe aceast\u0103 tem\u0103. <\/li>\n<li>Tehnologiile oferite de companiile Intel \u0219i AMD pentru monitorizarea RAM-ului \u0219i a cache-ului permit studierea comportamentului ma\u0219inilor virtuale \u0219i plasarea acestora astfel \u00eenc\u00e2t vecinii \"zgomoto\u0219i\" s\u0103 nu interfereze cu func\u021bionarea ma\u0219inilor virtuale \"lini\u0219tite\". <\/li>\n<li>Extinderea setului de parametri (re\u021bea, stocare, prioritatea ma\u0219inii virtuale, costul migra\u021biei, preg\u0103tirea pentru migrare).<\/li>\n<\/ol>\n<p><\/p>\n<h2>\u00cen concluzie<\/h2>\n<p>\nRezultatul muncii noastre \u00een \u00eembun\u0103t\u0103\u021birea algoritmilor de balansare este concluzia clar\u0103 c\u0103, datorit\u0103 algoritmilor moderni, se poate ob\u021bine o optimizare semnificativ\u0103 a resurselor (25-30%) \u00een centrele de date \u0219i, \u00een acela\u0219i timp, \u00eembun\u0103t\u0103\u021birea calit\u0103\u021bii serviciilor pentru clien\u021bi.<\/p>\n<p>Algoritmul bazat pe re\u021bele neuronale este, f\u0103r\u0103 \u00eendoial\u0103, interesant, dar are nevoie de dezvolt\u0103ri ulterioare \u0219i, din cauza constr\u00e2ngerilor existente, nu este potrivit pentru rezolvarea acestui tip de probleme la volumele caracteristice pentru cloud-urile private. \u00cen schimb, \u00een cloud-urile publice de dimensiuni considerabile, algoritmul a ar\u0103tat rezultate bune.<\/p>\n<p>Detalii despre capabilit\u0103\u021bile procesoarelor, planificatorilor \u0219i balans\u0103rii la nivel \u00eenalt le vom prezenta \u00een articolele urm\u0103toare.<br \/>\n<br \/>Sursa: <a content=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/rostelecom\/blog\/465715\/\">habr.com<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u0412 \u043f\u0440\u043e\u0448\u043b\u043e\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u0430\u043b\u0438 \u043e \u043f\u043e\u043f\u044b\u0442\u043a\u0430\u0445 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u044c Watcher \u0438 \u043f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u0438\u043b\u0438 \u043e\u0442\u0447\u0435\u0442 \u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0439. \u0422\u0430\u043a\u0438\u0435 \u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u044f \u043c\u044b \u043f\u0435\u0440\u0438\u043e\u0434\u0438\u0447\u0435\u0441\u043a\u0438 \u043f\u0440\u043e\u0432\u043e\u0434\u0438\u043c \u0434\u043b\u044f \u0431\u0430\u043b\u0430\u043d\u0441\u0438\u0440\u043e\u0432\u043a\u0438 \u0438 \u0434\u0440\u0443\u0433\u0438\u0445 \u043a\u0440\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0438\u0445 \u0444\u0443\u043d\u043a\u0446\u0438\u0439 \u0431\u043e\u043b\u044c\u0448\u043e\u0433\u043e \u043a\u043e\u0440\u043f\u043e\u0440\u0430\u0442\u0438\u0432\u043d\u043e\u0433\u043e \u0438\u043b\u0438 \u043e\u043f\u0435\u0440\u0430\u0442\u043e\u0440\u0441\u043a\u043e\u0433\u043e \u043e\u0431\u043b\u0430\u043a\u0430. \u0412\u044b\u0441\u043e\u043a\u0430\u044f \u0441\u043b\u043e\u0436\u043d\u043e\u0441\u0442\u044c \u0440\u0435\u0448\u0430\u0435\u043c\u043e\u0439 \u0437\u0430\u0434\u0430\u0447\u0438, \u0432\u043e\u0437\u043c\u043e\u0436\u043d\u043e, \u043f\u043e\u0442\u0440\u0435\u0431\u0443\u0435\u0442 \u043d\u0435\u0441\u043a\u043e\u043b\u044c\u043a\u0438\u0445 \u0441\u0442\u0430\u0442\u0435\u0439 \u0434\u043b\u044f \u043e\u043f\u0438\u0441\u0430\u043d\u0438\u044f \u043d\u0430\u0448\u0435\u0433\u043e \u043f\u0440\u043e\u0435\u043a\u0442\u0430. \u0421\u0435\u0433\u043e\u0434\u043d\u044f \u043c\u044b \u043f\u0443\u0431\u043b\u0438\u043a\u0443\u0435\u043c \u0432\u0442\u043e\u0440\u0443\u044e \u0441\u0442\u0430\u0442\u044c\u044e \u0446\u0438\u043a\u043b\u0430, \u043f\u043e\u0441\u0432\u044f\u0449\u0435\u043d\u043d\u0443\u044e \u0431\u0430\u043b\u0430\u043d\u0441\u0438\u0440\u043e\u0432\u043a\u0435 \u0432\u0438\u0440\u0442\u0443\u0430\u043b\u044c\u043d\u044b\u0445 \u043c\u0430\u0448\u0438\u043d \u0432 \u043e\u0431\u043b\u0430\u043a\u0435. \u041d\u0435\u043c\u043d\u043e\u0433\u043e \u0442\u0435\u0440\u043c\u0438\u043d\u043e\u043b\u043e\u0433\u0438\u0438 [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":28330,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[688],"tags":[],"class_list":["post-37745","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-administrirovanie"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.2.1 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u0412 \u043f\u0440\u043e\u0448\u043b\u043e\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u0430\u043b\u0438 \u043e \u043f\u043e\u043f\u044b\u0442\u043a\u0430\u0445 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u044c Watcher \u0438 \u043f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u0438\u043b\u0438 \u043e\u0442\u0447\u0435\u0442 \u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0439.\" \/>\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\/ro\/blog\/administrirovanie\/balansirovka-nagruzki-v-openstack-chast-2\" \/>\n\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.2.1\" \/>\n\t\t<meta property=\"og:locale\" content=\"ro_RO\" \/>\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\u0411\u0430\u043b\u0430\u043d\u0441\u0438\u0440\u043e\u0432\u043a\u0430 \u043d\u0430\u0433\u0440\u0443\u0437\u043a\u0438 \u0432 Openstack (\u0427\u0430\u0441\u0442\u044c 2) | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u0412 \u043f\u0440\u043e\u0448\u043b\u043e\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u0430\u043b\u0438 \u043e \u043f\u043e\u043f\u044b\u0442\u043a\u0430\u0445 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u044c Watcher \u0438 \u043f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u0438\u043b\u0438 \u043e\u0442\u0447\u0435\u0442 \u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0439.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/ro\/blog\/administrirovanie\/balansirovka-nagruzki-v-openstack-chast-2\" \/>\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=\"2019-10-31T19:19:28+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2019-10-31T19:19:28+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\udd47Balansarea \u00eenc\u0103rc\u0103rii \u00een Openstack (Partea 2) | ProHoster","description":"\u00cen articolul anterior, am discutat despre \u00eencerc\u0103rile de a utiliza Watcher \u0219i am prezentat raportul testelor.","canonical_url":"https:\/\/prohoster.info\/ro\/blog\/administrirovanie\/balansirovka-nagruzki-v-openstack-chast-2","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"ro_RO","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\u0411\u0430\u043b\u0430\u043d\u0441\u0438\u0440\u043e\u0432\u043a\u0430 \u043d\u0430\u0433\u0440\u0443\u0437\u043a\u0438 \u0432 Openstack (\u0427\u0430\u0441\u0442\u044c 2) | ProHoster","og:description":"\u0412 \u043f\u0440\u043e\u0448\u043b\u043e\u0439 \u0441\u0442\u0430\u0442\u044c\u0435 \u043c\u044b \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u0430\u043b\u0438 \u043e \u043f\u043e\u043f\u044b\u0442\u043a\u0430\u0445 \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u044c Watcher \u0438 \u043f\u0440\u0435\u0434\u0441\u0442\u0430\u0432\u0438\u043b\u0438 \u043e\u0442\u0447\u0435\u0442 \u0438\u0441\u043f\u044b\u0442\u0430\u043d\u0438\u0439.","og:url":"https:\/\/prohoster.info\/ro\/blog\/administrirovanie\/balansirovka-nagruzki-v-openstack-chast-2","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":"2019-10-31T19:19:28+00:00","article:modified_time":"2019-10-31T19:19:28+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"37745","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":"2026-01-23 19:08:19","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-03-01 01:21:23","updated":"2026-01-23 19:08:19","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts\/37745","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/comments?post=37745"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts\/37745\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/media\/28330"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/media?parent=37745"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/categories?post=37745"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/tags?post=37745"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}