ProHoster > Блог > ozi ịntanetị > Apache Cassandra 4.0 DBMS dị

Apache Cassandra 4.0 DBMS dị

Apache Software Foundation weputara ntọhapụ nke DBMS Apache Cassandra 4.0 ekesa, nke bụ nke klas nke sistemu noSQL ma emebere ya iji mepụta nnukwu nchekwa na ntụkwasị obi nke nnukwu data echekwara n'ụdị nhazi nhazi (hash). A na-ahụta ntọhapụ Cassandra 4.0 dị ka ọ dị njikere maka mmejuputa mmepụta yana anwalelarị ya na akụrụngwa nke Amazon, Apple, DataStax, Instaclustr, ala na Netflix nwere ụyọkọ nke ihe karịrị 1000 ọnụ. Edere koodu ọrụ ahụ na Java ma kesaa ya n'okpuru ikike Apache 2.0.

Facebook mebere Cassandra DBMS na 2009 bufee ya n'okpuru nkwado Apache Foundation. Ngwọta ụlọ ọrụ dabere na Cassandra ka etinyere na ọrụ ike sitere na ụlọ ọrụ dịka Apple, Adobe, CERN, Cisco, IBM, HP, Comcast, Disney, eBay, Huawei, Netflix, Sony, Rackspace, Reddit na Twitter. Dịka ọmụmaatụ, akụrụngwa nchekwa nke Apache Cassandra nke Apple wesara nwere ihe karịrị otu puku ụyọkọ, gụnyere ọnụ ọnụ 160 yana ịchekwa data karịrị 100 petabytes. Huawei na-eji ihe karịrị 300 Apache Cassandra ụyọkọ, gụnyere 30 puku ọnụ, na Netflix na-eji ihe karịrị 100 ụyọkọ, na-ekpuchi 10 puku ọnụ na nhazi ihe karịrị otu puku ijeri arịrịọ kwa ụbọchị.

Cassandra DBMS na-ejikọta sistemu hash Dynamo kesara nke ọma, nke na-enye ihe fọrọ nke nta ka ọ bụrụ scalability linear ka olu data na-abawanye. Cassandra na-eji ụdị nchekwa data dabere na ezinụlọ kọlụm (ColumnFamily), nke dị iche na sistemu dị ka memcachedb, nke na-echekwa data naanị na eriri igodo / uru, site n'ikike ịhazi nchekwa nke hashes nwere ọtụtụ ọkwa akwụ. Iji mee ka mmekọrịta dị mfe na nchekwa data, a na-akwado asụsụ ajụjụ a haziri ahazi CQL (Cassandra Query Language) nke na-echetara SQL, mana ọ na-ebelata na arụmọrụ. Atụmatụ gụnyere nkwado maka oghere aha na ezinụlọ kọlụm, yana imepụta ndeksi site na iji okwu "CREATE INDEX".

DBMS na-enye gị ohere ịmepụta nchekwa na-eguzogide ọdịda: data etinyere na nchekwa data na-emegharị na-akpaghị aka n'ọtụtụ ọnụ nke netwọk kesara, nke nwere ike gbasaa ebe data dị iche iche. Mgbe ọnụ na-ada, a na-eburu ọrụ ya na ijiji site na ọnụ ụzọ ndị ọzọ. Ịtinye ọnụ ọhụrụ na ụyọkọ na imelite ụdị Cassandra ka a na-eme na ofufe, na-enweghị enyemaka aka ọzọ ma ọ bụ nhazigharị ọnụ ndị ọzọ. A na-akwado ndị ọkwọ ụgbọala nwere nkwado CQL maka Python, Java (JDBC/DBAPI2), Ruby, PHP, C++ na JavaScript (Node.js).

Isi ihe ọhụrụ:

  • Emelitere arụmọrụ yana scalability. Ewelitela arụmọrụ mgbanwe data na usoro SSTable (Table Strings Table) n'etiti ọnụ. Edozila Protocol ozi Internode. Ọsọ nke ịnyefe iyi data n'etiti ọnụ amụbaala ruo ugboro 5 (karịsịa n'ihi iji usoro Zero Copy na mbufe SSTables dum), na ntinye maka ịgụ na ide ọrụ abawanyela na 25%. Emebere usoro mgbake na-abawanye. Akwụsịla nkwụsị n'ihi nkwụsịtụ mkpofu ahịhịa na-ebelata ka ọ bụrụ millisekọd ole na ole.
  • Nkwado agbakwunyere maka ndekọ nyocha nke na-enye gị ohere soro ọrụ njirimara njirimara yana ajụjụ CQL niile emezuru.
  • Agbakwunyere ikike idowe ndekọ arịrịọ ọnụọgụ abụọ zuru oke, na-enye gị ohere ịchekwa arịrịọ na nzaghachi nzaghachi niile. Maka njikwa, iwu "nodetool enablefullquerylog|disablefullquerylog|resetfullquerylog" ka atụpụtara, na e wetara fqltool ike maka nyocha ndekọ. A na-enye iwu maka ịtụgharị ndekọ ahụ n'ụdị enwere ike ịgụ (Ntupu), atụnyere mpekere ọrụ (Tụlee) na imegharị (Replay) maka nyocha na-emepụtaghachi ọnọdụ dị na ezigbo ibu.
  • Nkwado agbakwunyere maka tebụl mebere nke na-egosipụta na ọ bụghị data echekwara na SSTables, kama mpụta ozi sitere na API (metric arụmọrụ, ozi ntọala, ọdịnaya cache, ozi gbasara ndị ahịa ejikọrọ, wdg).
  • A na-emeziwanye arụmọrụ nke nchekwa data abịakọrọ, na-ebelata ohere diski na imeziwanye arụmọrụ ọgụgụ.
  • A na-etinye data metụtara igodo igodo sistemụ (sistemu.*) ugbu a na ndekọ mbụ na ndabara kama ekesa ya na akwụkwọ ndekọ aha data niile, nke na-enye ohere ka oghere ahụ nọgide na-arụ ọrụ ma ọ bụrụ na otu n'ime diski ndị ọzọ ada ada.
  • Nkwado nnwale agbakwunyere maka ntụgharị oge na-agafe agafe na ọnụ ọnụ ọnụ ala. Mpempe akwụkwọ nwa oge anaghị echekwa data niile wee jiri iweghachi agbakwunyere ka ọ dabara n'uju oyiri. Ọnụọgụ ọkụ na-arụ ọrụ ide njikarịcha nke a na-edoghị ede ka ọ bụrụ ihe nnochite nwa oge ruo mgbe enwere nhazi zuru ezu nke oyiri.
  • Nkwado nnwale agbakwunyere maka Java 11.
  • Nhọrọ nnwale agbakwunyere iji tụnyere Osisi Merkle niile. Dịka ọmụmaatụ, ịme nhọrọ na ụyọkọ ọnụ ọnụ 3 nke oyiri abụọ bụ otu na otu bụ stale ga-ebute mmelite nke stale oyiri site na iji naanị otu ọrụ oyiri nke ugbu a.
  • Agbakwunyere ọrụ ọhụrụ currentTimestamp, currentDate, currentTime na currentTimeUUID.
  • Nkwado agbakwunyere maka ọrụ mgbakọ na mwepụ na ajụjụ CQL.
  • Enyere ikike ịrụ ọrụ mgbakọ na mwepụ n'etiti data nwere ụdị "timestamp"/"ụbọchị" na "ogologo oge".
  • Agbakwunyere ọnọdụ maka ịhụchalụ iyi data achọrọ maka mgbake (nodetool mmezi —nhụchalụ) yana ikike ịlele iguzosi ike n'ezi ihe nke data eweghachiri (nodetool repair —validate).
  • Ajụjụ ahọpụtara ugbu a nwere ikike ịhazi Map na Tọọ ihe.
  • Nkwado agbakwunyere maka ime ka usoro ihe owuwu mbụ nke echiche ihe n'anya (cassandra.yaml:concurrent_materialized_view_builders).
  • Iwu "nodetool cfstats" agbakwunyela nkwado maka nhazi site na ụfọdụ metrik yana ịmachi ọnụọgụ nke ahịrị egosiri.
  • Enyere ntọala ka amachi njikọ onye ọrụ na ụfọdụ ebe data naanị.
  • Agbakwunyere ikike ịmachi ike (oke ọnụego) nke imepụta foto na mkpochapụ ọrụ.
  • cqlsh na cqlshlib na-akwado Python 3 (A ka na-akwado Python 2.7).
  • Akwụsịla nkwado maka ikpo okwu Windows. Iji na-agba Cassandra na Windows, a na-atụ aro ka iji gburugburu Linux emebere na ndabere nke WSL2 subsystem (Windows Subsystem for Linux 2) ma ọ bụ sistemụ arụmọrụ.



isi: opennet.ru

Tinye a comment