{"id":152005,"date":"2025-12-18T23:11:55","date_gmt":"2025-12-18T21:11:55","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/reliz-openzfs-2-4-0-realizaczii-zfs-dlya-linux-i-freebsd"},"modified":"2025-12-18T23:11:55","modified_gmt":"2025-12-18T21:11:55","slug":"reliz-openzfs-2-4-0-realizaczii-zfs-dlya-linux-i-freebsd","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/news\/reliz-openzfs-2-4-0-realizaczii-zfs-dlya-linux-i-freebsd","title":{"rendered":"L\u00ebshimi OpenZFS 2.4.0, realizimi i ZFS p\u00ebr Linux dhe FreeBSD","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Pas 11 muaj zhvillimi, \u00ebsht\u00eb publikuar l\u00ebshimi i projektit OpenZFS 2.4.0, q\u00eb zhvillon zbatimin e sistemit t\u00eb skedar\u00ebve ZFS p\u00ebr Linux dhe FreeBSD. Projekti u b\u00eb i njohur si \u201cZFS ne Linux\u201d dhe m\u00eb par\u00eb ishte i kufizuar n\u00eb zhvillimin e nj\u00eb moduli p\u00ebr b\u00ebrthamen e Linux, por pas bashkimit me kodin nga FreeBSD, u njoh si zbatimi kryesor i OpenZFS dhe u riem\u00ebrua. <\/p>\n
Punimi i OpenZFS \u00ebsht\u00eb verifikuar me b\u00ebrthamat Linux nga 4.18 deri n\u00eb 6.18 dhe t\u00eb gjitha deg\u00ebt e FreeBSD, duke filluar nga 13.3. Kodi shp\u00ebrndahet n\u00ebn licenc\u00ebn e lir\u00eb CDDL. OpenZFS tashm\u00eb p\u00ebrdoret n\u00eb FreeBSD dhe \u00ebsht\u00eb p\u00ebrfshir\u00eb n\u00eb distribucionet Debian, Ubuntu, Gentoo, NixOS dhe ALT Linux. Paketa me versionin e ri do t\u00eb p\u00ebrgatitet s\u00eb shpejti p\u00ebr distribucionet kryesore t\u00eb Linux, duke p\u00ebrfshir\u00eb Debian, Ubuntu, Fedora, RHEL\/CentOS. <\/p>\n
OpenZFS ofron nj\u00eb implementim t\u00eb komponent\u00ebve ZFS, t\u00eb lidhur si me funksionimin e sistemit t\u00eb skedar\u00ebve, ashtu edhe me funksionimin e menaxherit t\u00eb volumit. Jan\u00eb implementuar komponent\u00ebt: SPA (Storage Pool Allocator), DMU (Data Management Unit), ZVOL (ZFS Emulated Volume) dhe ZPL (ZFS POSIX Layer). Projekti gjithashtu lejon p\u00ebrdorimin e ZFS si backend p\u00ebr sistemin klaster t\u00eb skedar\u00ebve Lustre. P\u00ebrparimet e OpenZFS bazohen n\u00eb kodin origjinal ZFS, i importuar nga projekti OpenSolaris dhe i zgjeruar me p\u00ebrmir\u00ebsime dhe korrigjime nga komuniteti Illumos. Projekti zhvillohet n\u00eb bashk\u00ebpunim me punonj\u00ebsit e laboratorit komb\u00ebtar Livermore, sipas kontrat\u00ebs me Ministrin\u00eb e Energjis\u00eb t\u00eb SHBA. <\/p>\n
Kodi shp\u00ebrndahet n\u00ebn licenc\u00ebn e lir\u00eb CDDL, e cila \u00ebsht\u00eb e papajtueshme me GPLv2, \u00e7ka nuk lejon integrimin e OpenZFS n\u00eb deg\u00ebn kryesore t\u00eb b\u00ebrtham\u00ebs Linux, pasi p\u00ebrzierja e kodit n\u00ebn licencat GPLv2 dhe CDDL \u00ebsht\u00eb e papranueshme. P\u00ebr t\u00eb kap\u00ebrcyer papajtueshm\u00ebrin\u00eb e licencave, u vendos q\u00eb produkti p\u00ebr Linux t\u00eb shp\u00ebrndahet krejt\u00ebsisht n\u00ebn licenc\u00ebn CDDL si nj\u00eb modul i shkarkuesh\u00ebm ve\u00e7mas nga b\u00ebrthama. Stabiliteti i baz\u00ebs s\u00eb kodit OpenZFS vler\u00ebsohet si i krahasuesh\u00ebm me sistemet e tjera t\u00eb skedar\u00ebve p\u00ebr Linux. <\/p>\n
Ndryshimet kryesore: <\/p>\n
\n
\u00cbsht\u00eb realizuar mund\u00ebsia p\u00ebr t\u00eb vendosur kuota n\u00eb m\u00ebnyr\u00eb t\u00eb paracaktuar p\u00ebr p\u00ebrdoruesit, grupet dhe projektet. N\u00eb versionet e m\u00ebparshme, kuotat mund t\u00eb vendoseshin vet\u00ebm ve\u00e7mas p\u00ebr \u00e7do identifikues p\u00ebrdoruesi. N\u00eb versionin e ri \u00ebsht\u00eb shtuar mund\u00ebsia e caktimit t\u00eb nj\u00eb vler\u00eb t\u00eb paracaktuar, e cila do t\u00eb aplikohet, n\u00ebse p\u00ebr nj\u00eb p\u00ebrdorues t\u00eb caktuar nuk \u00ebsht\u00eb vendosur nj\u00eb kuot\u00eb.\n
\u00cbsht\u00eb shtuar mund\u00ebsia e kthimit t\u00eb mod\u00ebs s\u00eb input\/output direkt (O_DIRECT, Direct I\/O) n\u00eb p\u00ebrdorimin e mekanizmit t\u00eb leht\u00eb t\u00eb input\/output t\u00eb pa-cache (Uncached I\/O) p\u00ebr k\u00ebrkesat q\u00eb nuk jan\u00eb t\u00eb rregulluara sipas kufirit t\u00eb faqes s\u00eb memories, t\u00eb cilat nuk mb\u00ebshteten n\u00eb Direct I\/O.\n
\u00cbsht\u00eb propozuar nj\u00eb algorit\u00ebm i ri unifikues p\u00ebr kufizimin e shpejt\u00ebsis\u00eb s\u00eb operacioneve t\u00eb shkruajtjes dhe vendosjes s\u00eb blloqeve, i cili redukton fragmentimin e vdev (Virtual Device) p\u00ebrmes agregimit t\u00eb disa operacioneve t\u00eb shkruajtjes dhe p\u00ebrdorimit t\u00eb hap\u00ebsirave m\u00eb t\u00eb m\u00ebdha t\u00eb lira t\u00eb vazhdueshme p\u00ebr to, n\u00eb vend q\u00eb t\u00eb vendos\u00eb rezultatet e operacioneve t\u00eb ndara n\u00eb hap\u00ebsira t\u00eb vogla t\u00eb lira.\n
\u00cbsht\u00eb rritur performanca e enkriptimit p\u00ebr shkak t\u00eb p\u00ebrdorimit t\u00eb instrukcioneve AVX2 n\u00eb zbatimin e algoritmit AES-GCM. N\u00eb testet e kryera, p\u00ebrshpejtimi arrin deri n\u00eb 80%.\n
\u00cbsht\u00eb lejuar vendosja e blloqeve ZIL (ZFS Intent Log) jo vet\u00ebm n\u00eb vdev normal ose n\u00eb nj\u00eb pajisje t\u00eb ve\u00e7ant\u00eb SLOG, por edhe n\u00eb vdev t\u00eb ve\u00e7anta. N\u00ebse vdev normal \u00ebsht\u00eb i lidhur me nj\u00eb hard disk dhe vdev i ve\u00e7ant\u00eb me nj\u00eb SSD, kjo ndryshim lejon p\u00ebrjashtimin e situatave kur t\u00eb dh\u00ebnat shkruhen n\u00eb SSD, nd\u00ebrsa blloqet e lidhura ZIL n\u00eb hard disk, duke shkaktuar vonesa t\u00eb panevojshme.\n
\u00cbsht\u00eb lejuar caktimi i pron\u00ebs special_small_blocks p\u00ebr ndarjet (ZVOL), jo vet\u00ebm p\u00ebr sistemet e skedar\u00ebve, duke lejuar organizimin e vendosjes s\u00eb blloqeve t\u00eb vogla n\u00eb nj\u00eb vdev t\u00eb ve\u00e7ant\u00eb, p\u00ebr shembull, n\u00eb nj\u00eb SSD.\n
Shtuar m\u00ebnyra e riparimit \u201czfs rewrite -P\u201d, e cila, kur \u00ebsht\u00eb e mundur, e l\u00eb t\u00eb pandryshuar koh\u00ebn logjike t\u00eb krijimit, duke mund\u00ebsuar q\u00eb t\u00eb dallohen blloqet e thjeshta t\u00eb zhvendosura n\u00eb pool nga blloqet q\u00eb jan\u00eb ndryshuar me t\u00eb v\u00ebrtet\u00eb nga p\u00ebrdoruesi.\n
P\u00ebr komandat \u201czpool trim\u201d, \u201czpool scrub\u201d dhe \u201czpool initialize\u201d \u00ebsht\u00eb implementuar opsioni \u201c-a\u201d (\u2014all), i cili lejon q\u00eb k\u00ebto komanda t\u00eb ekzekutohen menj\u00ebher\u00eb p\u00ebr t\u00eb gjitha pullet e importuara.\n
Shtuar komanda \u201czpool scrub -S -E\u201d p\u00ebr t\u00eb kryer verifikimin e kontrolleve t\u00eb shumave t\u00eb t\u00eb dh\u00ebnave q\u00eb bien n\u00eb intervalet e caktuara t\u00eb koh\u00ebs.\n
Jan\u00eb shtuar optimizime t\u00eb reja p\u00ebr procesin e deduplication. <\/ul>\n