Aikin OpenZFS 2.2.1, wanda ke haɓaka aiwatar da tsarin fayil ɗin ZFS don Linux da kuma FreeBSD. Wannan fitowar ta shahara saboda ƙarin tallafin kernel. Linux 6.6 da kuma ƙoƙarin kawar da matsalar da ke haifar da lalacewar bayanai (rage wasu tubalan) a cikin fayiloli bayan an kwafi su.
Da farko, an yi imanin batun ya keɓanta da reshe na 2.2.x kuma ya haifar da bug a cikin tsarin toshe cloning wanda aka haɗa a cikin OpenZFS 2.2.0. Wannan tsarin yana ba da damar ƙirƙirar kwafin fayil ko ɓangarensa ba tare da kwafin bayanai ba, ta amfani da nassoshi na toshe bayanan da ke cikin ainihin fayil ɗin a cikin kwafi na biyu ba tare da kwafin su ba. A cikin OpenZFS 2.2.1, toshe cloning an kashe ta tsohuwa don rage matsalar, kuma an ƙara saitin zfs_bclone_enabled don maido da goyan bayan wannan yanayin.
Daga baya, masu haɓaka sun ba da rahoton sake buga matsalar a cikin tsari tare da reshen OpenZFS 2.1.x. Ba a tabbatar da zato cewa matsalar ta bayyana kanta a kan tsarin da ke da tsoffin fitowar fakitin coreutils ba - an sake buga kuskuren cikin nasara a cikin FreeBSD kuma Linux-rarrawa tare da sabon fitowar coreutils 9.4.
Lalacewar fayil yana faruwa a ƙarƙashin ingantattun yanayi. Misali, gudanar da umarnin "fito-1 dev-lang/go" akan sakamakon Gentoo a cikin shigar da kayan aikin Go, wanda ke lalata fayiloli a cikin /usr/lib/go/pkg/tool/linux_amd64/compile directory. An yi imanin cewa kuskuren ya fara bayyana kansa bayan saita siginar "zfs_dmu_offset_next_sync=1" ta tsohuwa a cikin openzfs 2.1.4. Har yanzu ba a gano tushen kuskuren ba. Saita ma'aunin "/sys/module/zfs/parameters/zfs_dmu_offset_next_sync" zuwa 0 shine shawarar da aka ba da shawarar.
source: budenet.ru
