RAID 5, RAID 6, ko ZFS Multiple Drive Drop: Farfado da Rushewar Magana, QNAP, ko TrueNAS Array Ba tare da Rusa Bambanci¶

Jin Synology DiskStation, QNAP TS-jerin, sabar TrueNAS, ko Linux mdadm Ma'ajiyar uwar garken ƙara ci gaba da ba da rahoto ** Crashed Volume **, ** Lalacewar Pool / Lalacewa ***, ko An cire haɗin faifai da yawa babban gaggawar ma'ajiya ce. Lokacin da tuƙi ɗaya ya faɗo daga a RAID 5, Synology Hybrid RAID (SHR-1), ko kuma ZFS RAIDZ1 Pool, tsararru tana gudana ba tare da sakewa ba (degraded state), sake gina tubalan da suka ɓace akan tashi ta hanyar karanta duk sauran faifan da suka rage a layi daya. Idan tuƙi na biyu ya faɗi a layi bayan sa'o'i, ko kuma ya jefa Kuskuren karantawa wanda ba'a iya murmurewa (URE) rabin gaba ta hanyar sake ginawa akan faifan maye gurbin, lissafin daidaito ya rushe kuma wurin ajiyar ajiya yana tafiya gaba daya. Masu kula da firgita sukan lalata terabytes na bayanan da za a iya dawo dasu a cikin dakika sittin masu zuwa ta danna Fara, sake ƙirƙira tsararru tare da mdadm --create, ko tilasta sake gina kan layi nan da nan akan kayan aikin da ya gaza. Idan katin mai sarrafa RAID na hardware da kansa ya mutu maimakon diski, canza zuwa Mutuwar katin RAID mai kula da shi, shigo da sifili na ƙasashen waje, kuma me yasa ba za ka taɓa fara faifai ba; idan rumbun kwamfutarka ɗaya yana danna ko ya ƙi yin juyi, tuntuɓi gazawar rumbun kwamfutarka da dawo da bayanai ko NAS RAID tsararrun ƙasƙanci, gazawar tuƙi da yawa, da dawo da ZFS/Btrfs. Idan wannan gazawar ajiya ta hana hoton likita na asibiti (PACS), aika gaggawar gaggawa, ko ababen more rayuwa-aminci, kira lambar gaggawa ta gida (911 a Amurka da Kanada, 112 a fadin Tarayyar Turai, 999 a Ƙasar Ingila) ko duba kundin adireshin mu na lambobin gaggawa.

Yi wannan yanzu¶

1. ** TSAYA kuma KAR a danna "Ƙaddamarwa," "Ƙirƙiri Pool," "Gyara," ko gudu mdadm --create a kan abubuwan da aka sauke ***: Danna Farawa a cikin Synology DSM ko QNAP QTS, ko gudu mdadm --create a cikin Linux, yana rubuta sabbin abubuwan toshewa kuma ya fara sifili da ratsi iri-iri a duk fayafai na memba. Tushen da aka sauke yawanci har yanzu yana riƙe 99.99% Fayilolin ku ba su da inganci, kawai an nuna ba a daidaita su ta wurin ma'aunin taron sa.
2. ** Ƙaddamar da NAS ko uwar garken da tsabta (ko cire wutar lantarki idan yana sake ginawa a cikin dannawa) don daskare abubuwan da ke faruwa ***: Kowane daƙiƙa tsararrun yana tsayawa kan layi a cikin wani yanki na ƙasƙanci, tsarin aiki yana ƙara babban toshewa. Update Time kuma Events counter akan faifai masu rai yayin da faifan da aka jefar ya tsaya a daskare a tsohuwar lambar taron. Idan NAS tana ƙoƙarin sake ginawa ta atomatik a cikin tuƙi tare da ɓangarori mara kyau, ci gaba da karanta wannan tuƙi na iya tuɓe murfin platter na Magnetic (head crash).
3. ** Lakabi kowane ramin tuƙi tare da alamar jiki ko tef (Bay 1, Bay 2, Bay 3, Bay 4) kafin zamewa kowane tire fita ***: Yayin Linux mdadm, LVM2, kuma ZFS odar tuƙi ta UUID a cikin kowane babban shingen yanki, tsofaffin masu sarrafa kayan masarufi da jagora dmsetup gyare-gyaren tebur yana buƙatar ainihin tsari na ramin jiki (0, 1, 2, 3...).
4. Kada kayi ƙoƙari murmurewa, tilastawa taro, ko fsck kai tsaye a kan tuƙi na asali na memba; clone kowane drive bit-for-bit ta amfani da ddrescue na farko: Kafin gudu mdadm --assemble --force ko zpool import -F, haɗa kowane faifan memba zuwa wurin aiki na Linux ta hanyar SATA/SAS kai tsaye (ba za a taɓa gadar USB mai arha da ke ɓoyewa ba. SMART ko SCT ERC timeouts) da kuma rufe shi da ddrescue kan lafiyayyun kayan aiki ko fayilolin hoto.
5. ** Tilasta-harhada tsararrun da aka kulle ta a cikin yanayin KARANTA-KAWAI (--readonly ko -o readonly=on***: Lokacin haduwa mdadm tsararraki waɗanda kididdigar taronsu suka ɓata a yayin faɗuwar ɗigon tuƙi guda ɗaya, koyaushe suna wucewa --readonly (mdadm --assemble --force --readonly /dev/md0 ...). A cikin yanayin karantawa kawai, ba haka ba mdadm, LVM2, ext4, Btrfs, kuma ba ZFS zai iya sake rubuta ratsi na daidaito ko sake kunna jarida mai lalata.
6. ** Kwafi mafi mahimmancin bayanan ku zuwa mashin ajiyar waje mai zaman kansa nan da nan da zarar an ɗora karatu-kawai ***: Fitar da takaddun ku, ma'ajin bayanai, injina, da hotuna zuwa keɓaɓɓen tuƙi da farko, sannan gogewa da gwada faifai daga karce.

Bishiyar yanke shawara: RAID mai tuƙi da yawa da gwajin gaggawa na tafkin ZFS¶

Yi amfani da wannan jadawali don tantance ko tsararrun ku sun sami raguwar ɓata lokaci mai ma'ana (TLER / ERC mismatch), gazawar ɓangarori masu yawa na gaskiya yayin sake ginawa, ko ƙungiyar ma'amala ta ZFS (TXG) laifi:

flowchart TD
    Start["RAID 5 / RAID 6 / SHR / ZFS Pool ya Rushe Bayan Faɗin Drive 2+"] --> StopIO["Dakatar da Duk Sake Gina & Wutar Lantarki NAS Nan take"]
    StopIO --> Label["Lakabi Kowane Tireshin Tuƙi ta Lambar Bay (Bay 1 zuwa Bay N)"]
    Label --> ConnectLinux["Haɗa Direbobin Membobi zuwa Mai watsa shiri na Ceto Linux Ubuntu ta hanyar SATA kai tsaye / HBA"]

    ConnectLinux --> CheckSmart["Duba SMART & Metadata Karanta-kawai: smartctl -a / mdadm --examine / zdb -l"]
    CheckSmart --> PhysicalHealth{"Shin Wasu Direbobi suna Dannawa, ƙin Juya, ko Faduwa a layi a dmesg?"}

    PhysicalHealth -->|"Ee: Rashin Injin Injiniya / Shugaban kan 2+ Drive a cikin RAID 5"| Cleanroom["TSAYA ƘARAR faifai -> Aika zuwa Lab ɗin Maido da Bayanan Tsabtace"]
    PhysicalHealth -->|"A'a: Duk Direbobi suna jujjuyawa (Zasu iya samun sashe mara kyau ko raguwar lokaci)"| CloneAll["Clone Kowane Memba Drive Ta Amfani da GNU ddrescue (Gaba Gaba -> Reverse Scrape)"]

    CloneAll --> DisconnectOrig["Cire Haɗin Direbobin Asali & Yi Aiki Tsaye akan Fayilolin Cloned / Hotuna"]
    DisconnectOrig --> ArrayType{"Wanne Gine-ginen Ajiya Aka Yi Amfani?"}

    ArrayType -->|"Synology DSM (SHR / Classic RAID)"| SynologyFlow["Duba Abubuwan Taɗi ta hanyar mdadm --examine -> Ƙarfafa-Tara Karatu-Kawai -> Kunna LVM2 / Btrfs Karanta-Kawai"]
    ArrayType -->|"QNAP QTS (HAL LVM bakin ciki / kauri RAID)"| QnapFlow["Haɗa mdadm Karanta-Kawai -> Kunna DRBD Layer idan Ya Gabatar -> Binciken Custom QNAP LVM2 Superblocks"]
    ArrayType -->|"TrueNAS / ZFS (RAIDZ1 / RAIDZ2 / Mirror)"| ZfsFlow["Gudanar da shigo da zpool -d / dev/disk/by-id -> Shigo da karantawa-kawai tare da Rollback: shigo da zpool -f -o readonly=on -F"]
    ArrayType -->|"Standard Linux mdam"| LinuxFlow["Kwatanta Ma'auni na Abubuwan -> Gudun mdadm --haɗa --force --readonly -> Dutsen Fayil tare da -o ro,noload"]

Mataki na 1: Me yasa tilasta sake ginawa yana lalata bayanai (The URE da Same-Batch Cascade)¶

Don fahimtar dalilin da yasa danna ** Gyara / Sake Gina *** ko maye gurbin tuƙi ɗaya da jira awanni 36 galibi yana haifar da faɗuwar tsararru, dole ne ku fahimci yadda sake fasalin daidaitawa ke jaddada tsufa na kafofin watsa labarai na maganadisu da kuma dalilin da yasa kwamfutocin tebur na mabukaci ke fita daga wuraren NAS.

1. Gaskiyar ilimin lissafi na Kurakurai da ba a iya dawo da su ba.UREs) yayin sake gina RAID 5 ko RAIDZ1¶

A cikin 4-drive RAID 5 ko ZFS RAIDZ1 Pool, data da kuma nau'ikan tubalan (P) an zare su a duk fayafai guda huɗu ta amfani da keɓaɓɓen OR (XOR) lissafi (D1 XOR D2 XOR D3 = P). Idan Drive 2 ya mutu gaba ɗaya, tsararrun ba ta da sauran jan aiki. Don sake gina ** tuƙi 2 *** akan sabon faifan maye gurbin, NAS dole ne ta karanta ** kowane yanki ɗaya daga byte 0 zuwa ƙarshen faifai ** akan ** tuƙi 1 **, ** tuƙi 3 *, da ** tuƙi 4 lokaci guda (Reconstructed D2 = D1 XOR D3 XOR P).

Ka yi la'akari da abin da ya faru lokacin da waɗancan fayafai ke da rumbun kwamfyutoci huɗu na terabyte 12 (gigabyte 12,000) waɗanda aka saya daga rukunin masana'anta iri ɗaya shekaru huɗu da suka gabata: - Sake gina terabytes 12 yana buƙatar karatu 36 terabytes (2.88 x 1014 guda rago) ci gaba da sama da sa'o'i 24 zuwa 48 a matsakaicin matsakaicin zafi da na inji. - Hard Drives na masu amfani yawanci suna ɗauke da Kuskuren karantawa wanda ba a iya dawo da shi ba (URE) bayani dalla-dalla 1 error in 1014 bits (12.5 terabytes), kuma abubuwan motsa jiki suna haɓaka kurakuran sassan ɓoyayyiyar shiru (Current_Pending_Sector) a cikin wuraren sanyi na faifan da ba a karanta ba a cikin shekaru. - Lokacin da Drive 3 ya ci karo da ɓangarori guda 4 da ba za a iya karantawa ba (0.004 megabyte) yayin da Drive 2 ya ɓace ko an sake gina shi, XOR Equation yana da abubuwa guda biyu da ba a sani ba**D2 kuma D3).

2. Me ya sa tebur tafiyarwa ba tare da TLER / SCT ERC a kori gaba ɗaya sama da sashe mara kyau¶

Anan ga tarkon da ke haifar da faɗuwar faɗuwar faɗuwar faɗuwar faɗuwar abubuwa akan Synology, QNAP, da Linux mdadm:

+-------------------------------------------------------------------------+
| WHY A SINGLE BAD SECTOR DROPS AN ENTIRE DRIVE DURING REBUILD            |
|                                                                         |
| 1. Drive 2 is dead/rebuilding. NAS reads Sector 849,102,400 on Drive 3. |
| 2. Drive 3 hits a weak magnetic spot (Current_Pending_Sector).          |
| 3. Consumer Drive Behavior (No TLER / SCT ERC):                         |
|    - Drive 3 enters deep internal error recovery for 30 to 120 seconds, |
|      stopping all SATA responses while retrying head positioning.       |
| 4. Linux Kernel / NAS Controller Timeout (Default 7 to 30 seconds):     |
|    - Kernel sees Drive 3 unresponsive for >30s -> Issues SATA bus reset |
|    - mdadm assumes Drive 3 is COMPLETELY DEAD and kicks it from array   |
| 5. RESULT: Both Drive 2 and Drive 3 are now marked FAULTY/REMOVED.      |
|    The entire RAID 5 array crashes offline immediately.                 |
+-------------------------------------------------------------------------+

Ka lura da ainihin abin da ya faru: Drive 3 bai mutu ba. Yana da 99.999999% sassan lafiya kuma kawai 1 ko 2 sassan kilobyte 4 da ba za a iya karantawa ba (8 kilobytes daga 12 terabytes). Duk da haka saboda mdadm ya kori Drive 3 daga cikin tsararru saboda ƙarewar umarni na daƙiƙa 30, gabaɗayan ƙarar yana nuna kamar Crashed*. Idan kun haɗa Drive 3 tare da ddrescue (wanda da kyau ya tsallake kilobytes 8 na ɓangarori marasa kyau kuma yana kwafin sauran terabytes 12) kuma ya tilasta-harra tsararrun karantawa kawai, kuna dawo da kusan kashi 100 na fayilolinku.

3. Me ya sa ba za ku taɓa yin takara ba mdadm --create ko danna "Initialize"¶

Lokacin da tsararru ta faɗo, tatsuniyoyi biyu masu lalacewa suna yawo akan layi: - ** Labari na 1: "Sake ƙirƙiri tsararru tare da ainihin odar tuƙi ta amfani da shi mdadm --create"** Ko da tare da --assume-clean, sigar zamani na mdadm tsoho zuwa daban-daban Data Offset dabi'u dangane da kwaya da mdadm saki (2048 sectors vs 262144 sectors). Yaushe mdadm --create ya rubuta sabon sigar 1.2 superblock a cikin kuskure mara kyau ko tare da girman guntun da bai dace ba, yana sake rubutawa LVM2 jigon ƙarar jiki ko ext4 superblock a gaban ƙarar. - Tatsuniya ta 2: "Fitar da mota na biyu da ya sauke, mayar da shi, sannan danna Gyara." Idan Drive 1 ya fita kwanaki uku da suka gabata (don haka bayanan sa sun cika kwanaki 3) kuma Drive 3 ya bar minti 10 da suka gabata (don haka bayanansa na yanzu), jawowa da sake shigar da motar da ba daidai ba ko tilasta sake ginawa tare da ** Drive 1 ** alamar "active" da ** Driver 3 ya sa aka sake rubutawa NPA tsararru tare da tsattsauran ratsi na kwana 3 **, lalata kowane kundin adireshi da fayil ɗin da aka gyara a cikin sa'o'i 72 da suka gabata.

Mataki na 2: Binciken superblocks da Event yana ƙididdigewa cikin aminci akan rundunar ceto ta Linux¶

Don kubutar da tsarin Synology, QNAP, TrueNAS, ko Linux lafiyayye, cire duk kayan aikin memba (bayan sanya lambobin bayyanin su) kuma haɗa su zuwa PC tebur ko sabar da ke gudana Ubuntu 22.04 LTS / 24.04 LTS (har ma an ɗaga su daga sandar USB ta Live).

Mataki 1: Shigar da kayan aikin ceto na forensic a cikin Ubuntu¶

Bude tasha akan injin ceto na Ubuntu kuma shigar da abubuwan da ake buƙata:

sudo apt-get update
sudo apt-get install -y gddrescue smartmontools mdadm lvm2 btrfs-progs xfsprogs zfsutils-linux

Mataki na 2: Taswirar kowace harafin tuƙi, lambar serial, da shimfidar ɓangarorin¶

Lissafin duk na'urorin toshewar da aka haɗa da lambobin kayan aikinsu don kada ku taɓa rikitar da faifan ceto tare da tushen diski NAS:

lsblk -o NAME,SIZE,MODEL,SERIAL,FSTYPE,TYPE,MOUNTPOINTS

Dubi teburin rarraba na al'ada ** Synology DSM ***, QNAP, ko TrueNAS tuƙi:

Mataki na 3: Duba mdadm --examine don ganin dalilin da yasa tsararrun ta fado (The Event Counter Matrix)¶

Kan Synology, QNAP, ko Linux mdadm tsararru, bincika metadata na RAID akan ɓangarori na bayanan mai amfani (misali, sashi na 3 akan /dev/sdb3, /dev/sdc3, /dev/sdd3, /dev/sde3) ba tare da hawa ko gyara komai ba:

sudo mdadm --examine /dev/sd[b-e]3 | grep -E "Filename|Magic|Update Time|Events|Array State|Device Role"

Kwatanta abin da aka fitar a cikin dukkan tukwici huɗu:

/dev/sdb3:
    Update Time : Sat Oct 31 14:12:09 2026
         Events : 482190
   Device Role : Active device 0
   Array State : A.AA ('A' == active, '.' == missing, 'R' == replacing)

/dev/sdc3:
    Update Time : Thu Oct 29 03:44:11 2026
         Events : 479012
   Device Role : Active device 1
   Array State : AAAA ('A' == active, '.' == missing, 'R' == replacing)

/dev/sdd3:
    Update Time : Sat Oct 31 14:12:04 2026
         Events : 482188
   Device Role : Active device 2
   Array State : A.AA ('A' == active, '.' == missing, 'R' == replacing)

/dev/sde3:
    Update Time : Sat Oct 31 14:12:09 2026
         Events : 482190
   Device Role : Active device 3
   Array State : A.AA ('A' == active, '.' == missing, 'R' == replacing)

Karanta wannan labarin na bincike a hankali: 1. /dev/sdc3 (Device Role 1* yana da Events ƙidaya 479012 daga kwana biyu da suka wuce (Oct 29). Wannan tuƙi ya ƙare da farko, kwanaki biyu da suka gabata, kuma tun daga lokacin ba a kan layi. 2. /dev/sdb3 (Role 0* da/dev/sde3 (Role 3 biyu suna da Events : 482190 a 14:12:09. 3. /dev/sdd3 (Role 2* yana da Events : 482188 a 14:12:04, kawai 5 seconds da 2 events baya* sdb3 kuma sde3. 4. **Kammalawa *: Dole ne ku sake gina tsararrun ta amfani da sdb3 (Role 0), sdd3 (Role 2), kuma sde3 (Role 3 (Wanda kididdigar taronsu takai ticks 2 kawai), yayin da ** barin sdc3 (Role 1) fita gaba daya.

Mataki na 3: Rufe kowane drive bit-for-bit tare da GNU ddrescue¶

Kafin gudu mdadm --assemble --force ko zpool import -F, clone kowane drive ɗin da ake buƙata (musamman mashin ɗin da ya bar ƙarshe saboda munanan sassa, kamar /dev/sdd a cikin misalin da ke sama) a kan lafiyayyen tuƙi mai ƙarfi na daidai ko girma, ko cikin ɗanye .img fayil akan babban ƙarar ajiya.

Mataki 1: Me yasa ddrescue tare da taswirar taswira wajibi ne¶

Daidaitawa dd zubar da ciki a kan mummunan sashe na farko, ko tare da conv=noerror,sync ɓata sa'o'i yana niƙa kan karantawa akan hanyar da ta lalace yayin da sifili-cike megabytes na kyawawan bayanai masu kyau. GNU ddrescue yana aiki a cikin ƙwararrun ƙwararrun matakai: 1. ** Wucewa 1 (lokacin kwafi): Kwafi duk lafiyayye, yankuna masu saurin amsawa na faifai a cikin cikakken saurin jeri.150 MB/s ku 250 MB/s), nan da nan yin tsalle a kan kowane shinge na jinkiri ko ba za a iya karantawa ba don kada motar ta yi zafi ko mutu kafin ka sami 99.99% data mai kyau. 2. ** Wuce 2 (lokacin gyarawa): Yana karantawa daidai gefuna na kowane mugun shinge daga gaba da baya (reverse direction) don ware yankin da ya lalace har zuwa ainihin sashin kayan aikin kilobyte 4. 3. Masu wuce 3 (lokacin zage-zage): Sake gwada ɓangarori da suka lalace a hankali ba tare da sake saita bas ɗin ba.

Mataki 2: Gudanar da mataki 2 ddrescue umarnin clone¶

Koyaushe saka ** mapfile** (sdd_rescue.map). Idan direban tushen da ke mutuwa ya katse kuma yana buƙatar sake zagayowar wuta, ddrescue yana karanta taswirar kuma ya ci gaba a daidai byte inda ya tsaya ba tare da sake karanta sashin da aka riga aka kwafi ba:

# Stage 1: Fast forward pass skipping bad/slow blocks immediately (-n = no-scrape)
sudo ddrescue -f -n -b 4096 /dev/sdd /dev/sdh /root/sdd_rescue.map

# Stage 2: Targeted trimming and 3-retry scrape of the skipped bad sectors
sudo ddrescue -f -d -r 3 -b 4096 /dev/sdd /dev/sdh /root/sdd_rescue.map

Mataki na 4: Sake Gina Synology DSM (SHR-1 / SHR-2 / Classic RAID), QNAP, da Linux mdadm¶

Da zarar na'urorin madauki na ku (ko na'urorin madauki-karanta kawai) sun shirya kuma an cire kayan aikin NAS na asali cikin aminci, bi ainihin jerin abubuwan da ba na lalacewa ba don dandalin ku.

Halin A: Standard Linux mdadm ko Synology Classic RAID (RAID 5 / RAID 6)¶

1. Dakatar da kowane bangare md tsararru cewa Ubuntu na iya ƙoƙarin farawa:

   cat /proc/mdstat
   sudo mdadm --stop /dev/md127
   

2. Ƙaddamar da ɓangarorin cloned (sdb3, sdd3, sde3, tsallake matattu sdc3) a cikin yanayin karantawa kawai*:

   sudo mdadm --assemble --force --readonly --verbose /dev/md2 /dev/sdb3 /dev/sdd3 /dev/sde3
   

   Kalli abin da aka fitar: mdadm zai bayar da rahoto forcing event count in /dev/sdd3(2) from 482188 up to 482190 kuma mdadm: /dev/md2 has been started with 3 drives (out of 4).
3. Tabbatar cewa tsararrun tana aiki da alama (auto-read-only) ko (read-only):

   cat /proc/mdstat
   

Yanayin B: Synology Hybrid RAID (SHR-1 kuma SHR-2) da LVM2 kuma Btrfs / ext4¶

Synology's ** SHR (Synology Hybrid RAID*** shine daidaitaccen Linux kashi 100 mdadm + LVM2 (Logical Volume Manager*: - Lokacin da duk abubuwan tuƙi a cikin tafkin SHR girmansu ɗaya ne (misali, fayafai 8 terabyte guda huɗu), Synology yana ƙirƙirar guda ɗaya. mdadm RAID 5 bangare (md2 a partition 3 na kowane drive) da kuma sanya wani LVM2 rukunin girma (vg1000 ko vg1) a saman md2. - Lokacin da tuki a cikin tafkin SHR yana da girma dabam (misali, terabyte 8 na terabyte guda biyu da terabyte 16 guda biyu), Synology yana ƙirƙirar daban daban. mdadm tsararraki: md2 Tsakanin Partition 5 (8 TB yanki a kan duk 4 drives a RAID 5) da md3 fadin Partition 6 (sauran 8 TB yanki a kan faifan TB guda 16 a cikin RAID 1), sannan a haɗa md2 + md3 cikin guda daya LVM2 Rukunin Juzu'i (vg1000).

Don hawa ƙarar ƙarar SHR na Synology akan Ubuntu bayan haɗuwa-ƙarfi /dev/md2 (kuma /dev/md3 idan gauraye-size SHR): 1. Bincika don Synology LVM2 Ƙirar Jiki da Ƙungiyoyin Ƙarfafa:

sudo pvscan
sudo vgscan
sudo lvscan

sudo vgchange -ay /dev/vg1000

sudo mkdir -p /mnt/synology_rescue
sudo mount -t ext4 -o ro,noload /dev/vg1000/lv /mnt/synology_rescue

sudo mkdir -p /mnt/synology_rescue
sudo mount -t btrfs -o ro,norecovery /dev/vg1000/lv /mnt/synology_rescue

sudo mount -t btrfs -o ro,norecovery,usebackuproot /dev/vg1000/lv /mnt/synology_rescue

Scenario C: Synology DSM 7+ Encryption Volume ko Shared Jaka boye-boye (eCryptfs)¶

• ** Ruɓan babban fayil ɗin da aka raba (@sharename@ manyan fayiloli dauke da ECRYPTFS_FNEK_ENCRYPTED... filenames)**: Kwafi duka @encrypted_share@ kashe babban fayil /mnt/synology_rescue da farko, sannan ku matsa ku yanke shi akan Ubuntu ta amfani da kalmar sirrin ɓoyewar Synology ko .key fayil:

  sudo apt-get install -y ecryptfs-utils
  sudo mkdir -p /mnt/decrypted_share
  sudo mount -t ecryptfs -o ro /mnt/synology_rescue/@encrypted_share@ /mnt/decrypted_share
  

• ** DSM 7.2+ Cikakken Rubutun Rubutun (LUKS***: Idan /dev/vg1000/lv nuna TYPE="crypto_LUKS" in blkid, buše shi karanta-kawai ta amfani da .rkey An sauke fayil ɗin Maɓalli na farfadowa lokacin da aka ƙirƙiri ƙarar:

  sudo cryptsetup open --readonly --key-file /path/to/Volume1.rkey /dev/vg1000/lv synology_crypt
  sudo mount -o ro,norecovery /dev/mapper/synology_crypt /mnt/synology_rescue
  

Yanayin D: QNAP QTS (LVM2 Thin Provisioning kuma DRBD Layer)¶

Rukunin QNAP NAS na zamani (QTS 4.3+ / 5.x) saka kumburin ciki guda ɗaya DRBD na'ura (/dev/drbd1) tsakanin /dev/md1 da kuma LVM2 Rukunin Juzu'i (vg1). 1. Da farko tilasta-harhada sassan bayanan QNAP (/dev/sd[b-e]3) karanta-kawai cikin /dev/md1:

sudo mdadm --assemble --force --readonly /dev/md1 /dev/sdb3 /dev/sdd3 /dev/sde3

Mataki na 5: Ceto TrueNAS da Linux ZFS wuraren waha (RAIDZ1, RAIDZ2, RAIDZ3)¶

Idan uwar garken ajiyar ku yana gudana ** TrueNAS CORE , TrueNAS SCALE, Proxmox VE, ko Ubuntu ZFS, dawo da baya amfani. mdadm. Madadin haka, ZFS yana haɗa mai sarrafa ƙarar da kwafi-kan-rubuta (CoW) tsarin fayil a cikin gine-ginen ma'amala guda ɗaya da aka gina a kusa da Uberblocks da jerin * ƙungiyoyin ma'amala.TXGs).

Mataki 1: Bincika alamun 4 ZFS vdev (L0, L1, L2, L3) akan kowane faifan cloned¶

Kowane ɓangaren memba na ZFS yana adana kwafi iri ɗaya guda huɗu na daidaitawar tafkin da tsararrun 128 na ƙarshe ** Uberblocks *** (alamomi biyu a gaban ɓangaren, L0 kuma L1, da kuma tambari biyu a ƙarshensa, L2 kuma L3). Gudu zdb -l akan kowane bangare na cloned don duba yanayin tafkin da kuma mafi girman ƙungiyar Ma'amala (txg):

sudo zdb -l /dev/sdb2 | grep -E "name:|state:|txg:|guid:" | head -n 20

Kwatanta da txg: lambobi a duk fa'idodin cloned. Kamar dai mdadm's Events counter, idan faifan diski ɗaya ya bar minti 5 kafin asarar wuta ko faɗuwar faifai na biyu, ta txg zai zama ƙasa kaɗan fiye da sauran faifai.

Mataki 2: Shigo da tafkin ZFS a cikin tsayayyen yanayin Karanta-kawai (-o readonly=on)¶

Da farko duba wuraren tafki masu shigo da kaya ba tare da hawa komai ta hanyar ba sudo zpool import -d /dev/disk/by-id. Kada a taɓa shigo da tafkin ZFS mai lalacewa ko faifai da yawa a cikin yanayin karantawa. Lokacin shigo da shi tare da -o readonly=on, ZFS ba ya ƙoƙarin ƙaddamar da resilver, baya rubuta sabon Uberblocks, kuma ba zai iya firgita kernel ba idan taswirar sararin samaniya ta lalace (spacemap) an ci karo da shi yayin rabon rubutu:

sudo mkdir -p /mnt/zfs_rescue
sudo zpool import -f -o readonly=on -R /mnt/zfs_rescue tank

Mataki na 3: Idan zpool import kasa da I/O error ko corrupted data, Yi amfani da TXG Rollback (-F kuma -T)¶

Idan faifai guda biyu sun fita a cikin daƙiƙa na juna yayin fashewar rubutu mai nauyi, ƙungiyar ma'amala ta ƙarshe (TXG N) na iya zama bai cika ba a duk faɗin ratsin da ke tsira, yayin da TXG N-1 ko TXG N-2 (an rubuta 5 zuwa 15 seconds a baya) kashi 100 cikin 100 cikin 100 cikin 100 na buffer zoben Uberblock.

1. ** Gwada samfoti na baya-bayan nan mara lalacewa (-F -n***:

   sudo zpool import -f -o readonly=on -R /mnt/zfs_rescue -F -n tank
   

   Idan ZFS tayi rahoto Would be able to return pool to its state as of <timestamp>, gudanar da shigo da koma baya na karanta-kawai ba tare da -n:

   sudo zpool import -f -o readonly=on -R /mnt/zfs_rescue -F tank
   

2. Idan -F har yanzu ba zai iya hawa sabbin Uberblocks ba, jera lambobi na Uberblock TXG na tarihi tare da zdb kuma shigo da takamaiman tsohuwar TXG (-T*:

   sudo zdb -e -ul /dev/sdb2 | grep -E "Uberblock|txg =|timestamp ="
   

   Zaba a txg lamba daga 30 zuwa 60 seconds kafin hadarin (misali, txg = 9481020) kuma shigo da tafkin da aka daskare a daidai lokacin a cikin yanayin karantawa kawai:

   sudo zpool import -f -o readonly=on -R /mnt/zfs_rescue -T 9481020 tank
   

Mataki na 6: Tabbacin dawowa bayan dawowa da hana sake gina tarkace nan gaba¶

Da zarar ka kwafi kashi 100 na bayanan da aka kwato daga /mnt/synology_rescue ko /mnt/zfs_rescue a kan maƙasudin madadin mai zaman kansa, bi wannan kayan aikin da lissafin daidaitawa kafin mayar da NAS ɗin ku cikin sabis:

• **Kada ku taɓa siyan duk kayan aikin NAS daga dillali ɗaya a rana guda ***: Masu tuƙi daga masana'anta iri ɗaya suna raba batches iri ɗaya da firmware iri ɗaya, yana sa su gaza cikin sa'o'i na juna yayin sake ginawa. Mix batches ko masana'anta (Seagate IronWolf / Exos, WD Red Plus / Ultrastar, Toshiba N300 / MG).
• ** Yi amfani da kayan aikin NAS/Kasuwanci tare da TLER / SCT ERC kunna (ko haɓaka Linux mdadm Lokacin SCSI don faifan tebur)**: Don abubuwan tafiyar da mabukata a cikin Linux mdadm, amfani a udev mulki don tadawa /sys/block/sdX/device/timeout daga 30 dakika zuwa 180 dakiku haka mdadm yana jiran sake taswirar ɓangaren cikin gida maimakon faɗuwar tuƙi.
• Yi ƙaura tare da faifai mafi girma fiye da 8 TB daga RAID 5 / SHR-1 / RAIDZ1 ku RAID 6 / SHR-2 / RAIDZ2: Da 12 TB ku 24 TB tafiyarwa, lokutan sake gina ɗaiɗai ɗaya sun wuce sa'o'i 30, suna yin dual-pararity (RAID 6 / SHR-2 / RAIDZ2) yana da mahimmanci don tsira daga mummunan sashe a lokacin sake dawowa.
• ** Jadawalin kowane wata Data Scrubbing (mdadm check / zpool scrub) da kuma Extended SMART gwaje-gwaje ***: Gyaran gogewa kowane wata yana gyara kurakurai a ɓangarori (Current_Pending_Sector) yayin da duk direbobi suna kan layi.
• Ka tuna cewa RAID babban samuwa ne, ba a taɓa samun ajiyar kuɗi ba: Bi 3-2-1 dabarun madadin tare da kwafin layi.

Lambobin gaggawa na yanki da ka'idojin hukumar¶

Kashewar ajiya da ke shafar cibiyoyin sadarwar kiwon lafiya, sadarwar gaggawa, ko bayanan sirri na doka suna buƙatar haɓaka nan take:

• Amurka da Kanada: Dial 911 idan hadarin ajiya ya hana tsarin asibiti na asibiti (EHR / PACS) ko aikawa. Rahoton NAS da aka yi niyya ransomware (DeadBolt, LockBit, Akirada CISA (1-844-Say-CISA / cisa.gov/report, FBI IC3 (ic3.gov, ko ** Cibiyar Tsaro ta Intanet ta Kanada (Cibiyar Tsaro ta Intanet)1-833-CYBER-88)**.
• ** United Kingdom **: Dial 999 don al'amuran gaggawa da suka shafi amincin rayuwa ko m kulawar NHS. A karkashin UK GDPR, tantance ko asarar bayanan sirri mara tallafi na dindindin yana buƙatar sanar da ICO (0303 123 1113 a cikin awanni 72.
• Ƙungiyar Tarayyar Turai: Dial112* a cikin EU. Hukumomin da ke ƙarƙashin NIS2 Directive dole ne su bayar da rahoton manyan abubuwan da aka dakatar da aiki ga CSIRT na ƙasa a cikin sa'o'i 24, da asarar wadatar bayanan sirri na dindindin ** Rahoton GDPR Mataki na 33** cikin sa'o'i 72.
• ** Ostiraliya da New Zealand : Kiran kira 000 a Ostiraliya ko 111 a New Zealand; bayar da rahoton abubuwan more rayuwa ko abubuwan da suka faru na ransomware NAS ga ACSC (1300 CYBER1 ko ** CERT NZ (0800 237 869).
• **Indiya da Asiya-Pacific **: Kiran kira 112 a Indiya don amsa gaggawa; bayar da rahoton abubuwan ajiyar kayan more rayuwa zuwa CERT-In (incident@cert-in.org.in a cikin taga tilas na awa 6.

Tambayoyin da ake yawan yi¶

Me ya sa mdadm --assemble --force kin fara RAID 5 na da not enough to start the array?¶

A cikin 4-drive RAID 5 tsararru, a kalla 3 cikin 4 tuƙi dole ne a wuce zuwa mdadm --haɗa --ƙarfi. Gudu sudo mdadm --examine /dev/sd[a-z]3 don bincika dalilin da yasa aka ƙi na'urar ta uku: ko dai kun wuce dukkan faifai maimakon bangare (/dev/sdd maimakon /dev/sdd3), an yiwa babban katangar sa alama a matsayin a Ajiye ta hanyar sake ginawa da aka katse, ko kuma na'urar ta atomatik ta Ubuntu ta kulle tuƙi ɗaya cikin mara aiki /dev/md127 na'ura (run sudo mdadm --stop --scan na farko, sa'an nan kuma sake gudu mdadm --haɗa --ƙarfi --karanta kawai).

Me zai faru idan motar ta fara sake ginawa (Recovery = 12%) kafin tuƙi na biyu ya fita?¶

Idan Drive 2 ya mutu, kun saka mai maye gurbin Drive 2, kuma a 12% sake gina ci gaba Drive 3 ya ragu a layi saboda mummunan yanki, kar a yi amfani da 12%-sake ginawa Drive 2. Madadin haka, clone na asali Turi 3 tare da damuwa. Saboda Drive 3 ya kasance gaba ɗaya a halin yanzu har sai an tsaya sake ginawa a kashi 12%, tilasta-harhada tsararrun karantawa-kawai ta amfani da Drive 1, damuwa clone na Drive 3, da Drive 4 (barin Drive 2 fita) yana ba ku daidaitaccen tsari.

Shin zan iya sanya kayan aikin Synology ko QNAP a cikin sabon chassis na NAS idan wutar lantarki ko hukumar NAS ta mutu?¶

Ee, duka Synology (Hijira HDDda QNAP (Hijira NAS kai tsaye) adana tsarin RAID akan abubuwan tafiyar da kansu. Duk da haka, idan guda biyu ko fiye sun riga sun fita (`Volume Crashed`) kafin chassis ya mutu, kar a sanya su a cikin sabon chassis na NAS, saboda mayen ƙaura na iya sake rubuta ɓangarori ko haifar da sake ginawa mai lalacewa. Ceto bayanan karanta-kawai akan Ubuntu farko.

Ta yaya zan iya sanin ko faɗuwar tuƙi yana da aminci don haɗawa da shi ddrescue sabanin bukatar dakin bincike mai tsafta?¶

Haɗa motar zuwa SATA, kunna shi, kuma saurare tsawon daƙiƙa 10 yayin kallo sudo dmesg -w. Idan drive ɗin yana yin latsa rhythmic, scraping, ko buzzing sautuna, jujjuya sama da jujjuyawa baya, ko yin rahoton ƙarfin kamar 0 B / 4.1 GB in lsblk, shugabannin karanta/rubutu sun gaza a zahiri: yanke wuta nan da nan kuma yi amfani da dakin gwaje-gwaje mai tsabta. Idan drive ɗin yana jujjuya sumul, yana nuna cikakken ƙarfinsa a ciki lsblk, da kuma gundumomi kawai Sashin_Yanzu ko kurakurai na lokaci, rufe shi da damuwa.

Menene "RAID 5 Write Hole," kuma ta yaya yake ZFS ko Btrfs kauce masa a lokacin da aka kashe wutar lantarki?¶

A gargajiyance RAID 5 / RAID 6 ba tare da mujallu ba, sabunta ɗigon yana buƙatar rubuta toshe bayanai da toshe toshe zuwa fayafai guda biyu a lokaci guda. Idan wutar ta gaza tsakiyar rubutawa, bayanai da daidaito ba za su ƙara daidaita ba (da Rubuta Hole). ZFS (`RAIDZ`) kuma Btrfs kawar da ramin rubutu ta amfani da Kwafi-kan-Rubuta (`CoW`), rubuta kowane sabon ratsin cikin tubalan da ba a yi amfani da su ba da kuma sabunta ma'anar Uberblock/tushen kawai bayan cikakken tsiri yana kan faifai.

Zan iya amfani dmsetup overlays don gwada sigogin RAID ba tare da gyaggyara abubuwan tafiyarwa ba?¶

Ee, zaku iya ƙirƙirar Kwafi-kan-Rubuta (`CoW`) overlays a saman na'urorin madauki kawai karantawa (`losetup --read-only`) ta amfani da Linux dmsetup (`snapshot` manufa) wanda ba shi da ƙarfi a cikin fayilolin da ba su da ƙarfi /tmp. Duk wani rubutu daga mdadm --haɗa --ƙarfi ko fsck -y shiga cikin abin zubarwa /tmp fayil mai rufi, yana ba ku damar sake saitawa a cikin daƙiƙa 1 idan an buƙata.

Jagora masu alaƙa¶

• NAS RAID tsararrun ƙasƙanci, gazawar tuƙi da yawa, da dawo da ZFS/Btrfs
• Mutuwar katin RAID mai kula da shi, shigo da sifili na ƙasashen waje, kuma me yasa ba za ka taɓa fara faifai ba
• Rashin gazawar rumbun kwamfutarka da dawo da bayanai
• Yanayin karantawa kawai na SSD da NVMe kwatsam, 0% katanga, da TRIM yana kashe ceto
• Hatsari rm -rf ko `Shift+Delete' akan Linux, macOS, ko Windows: nan da nan ya sake dawowa da karantawa kawai
• Kwamfuta madadin, dabarun 3-2-1, da dawo da bala'i
• Lambobin gaggawa ta ƙasa

Tushen da jagorar hukuma¶

• Linux Raid Wiki (raid.wiki.kernel.org) & Ƙungiya ta Linux Kernel: * Farfado da RAID na Software wanda bai yi nasara ba (mdadm --assemble --force --readonly da Matsalolin Matsalolin Matsala)*.
• Gidauniyar Software ta Kyauta (GNU Project): GNU ddrescue Manual: Farfajiyar Data farfadowa da na'ura daga Failing Media.
• Takardun Buɗe ZFS (openzfs.github.io): ZFS Pool Import farfadowa da na'ura (zpool import -F, Ana Shigo da Karatu kawai, da zdb Dubawa).
• Synology Inc.: * Cibiyar Ilimin Synology: Ta yaya zan iya amfani da PC don dawo da bayanai lokacin da Synology NAS ta yi rauni (mdadm + LVM2 Hawaye)*.
• Cibiyar Ƙididdiga da Fasaha ta Ƙasa (NIST) (Amurka): Bugawa na Musamman 800-209: Sharuɗɗan Tsaro don Kayayyakin Kayan Ajiye.