[LUGOS] RAID1 obcasne tezave

[Jurij Reščič]

>Jure Pecar wrote: Ce ti linux sw raid zabrise eno particijo iz arraya, vedno tudi napise razlog, 
>zakaj je to naredil. Podrobneje si oglej loge ... 
>
Evo, loge sem ogledal podrobneje, vendar ne vem, ali se čas napake ujema
s padcem arraya (vsaj zadnji zapis se zagotovo ne).  Je pa do tega 
najbrž prišlo
ob podobni situaciji.

Očitno sta oba diska precej slabega zdravja.

---------------------------------
Mar 18 14:45:41 master kernel: hdg: dma_intr: status=0x51 { DriveReady 
SeekComplete Error }
Mar 18 14:45:42 master kernel: hdg: dma_intr: error=0x40 { 
UncorrectableError }, LBAsect=241712, sector=37088
Mar 18 14:45:42 master kernel: end_request: I/O error, dev 22:02 (hdg), 
sector 37088
Mar 18 14:45:42 master kernel: raid1: Disk failure on hdg2, disabling 
device.
Mar 18 14:45:42 master kernel: ^IOperation continuing on 1 devices
Mar 18 14:45:42 master kernel: raid1: hdg2: rescheduling block 37088
Mar 18 14:45:42 master kernel: md: updating md3 RAID superblock on device
Mar 18 14:45:42 master kernel: md: hde2 [events: 00000042]<6>(write) 
hde2's sb offset: 40960000
Mar 18 14:45:42 master kernel: md: recovery thread got woken up ...
Mar 18 14:45:42 master kernel: md3: no spare disk to reconstruct array! 
-- continuing in degraded mode
Mar 18 14:45:42 master kernel: md: (skipping faulty hdg2 )
Mar 18 14:45:42 master kernel: raid1: hde2: redirecting sector 37088 to 
another mirror
------------------------------------------
Mar 18 14:45:51 master kernel: hde: dma_intr: status=0x51 { DriveReady 
SeekComplete Error }
Mar 18 14:45:51 master kernel: hde: dma_intr: error=0x40 { 
UncorrectableError }, LBAsect=249918, sector=45288
Mar 18 14:45:51 master kernel: end_request: I/O error, dev 21:02 (hde), 
sector 45288
Mar 18 14:45:51 master kernel: raid1: hde2: rescheduling block 45288
Mar 18 14:45:51 master kernel: raid1: hde2: unrecoverable I/O read error 
for block 45288
Mar 18 14:45:51 master kernel: journal_bmap_Rsmp_514801a1: journal block 
not found at offset 5132 on md(9,3)
Mar 18 14:45:51 master kernel: Aborting journal on device md(9,3).
Mar 18 14:45:51 master kernel: ext3_abort called.
Mar 18 14:45:51 master kernel: EXT3-fs abort (device md(9,3)): 
ext3_journal_start: Detected aborted journal
Mar 18 14:45:51 master kernel: Remounting filesystem read-only
Mar 18 14:45:51 master kernel: EXT3-fs error (device md(9,3)) in 
start_transaction: Journal has aborted
Mar 18 14:46:43 master last message repeated 4 times
Mar 18 14:46:44 master last message repeated 25 times
-----------------------------------------------
Apr  5 06:46:22 master kernel: hde: dma_intr: status=0x51 { DriveReady 
SeekComplete Error }
Apr  5 06:46:22 master kernel: hde: dma_intr: error=0x40 { 
UncorrectableError }, LBAsect=23365505, sector=23160880
Apr  5 06:46:22 master kernel: end_request: I/O error, dev 21:02 (hde), 
sector 23160880
Apr  5 06:46:22 master kernel: raid1: Disk failure on hde2, disabling 
device.
Apr  5 06:46:22 master kernel: ^IOperation continuing on 1 devices
Apr  5 06:46:22 master kernel: raid1: hde2: rescheduling block 23160880
Apr  5 06:46:22 master kernel: md: updating md3 RAID superblock on device
Apr  5 06:46:22 master kernel: md: (skipping faulty hde2 )
Apr  5 06:46:22 master kernel: md: hdg2 [events: 00000049]<6>(write) 
hdg2's sb offset: 40960000
Apr  5 06:46:22 master kernel: md: recovery thread got woken up ...
Apr  5 06:46:22 master kernel: md3: no spare disk to reconstruct array! 
-- continuing in degraded mode
Apr  5 06:46:22 master kernel: raid1: hdg2: redirecting sector 23160880 
to another mirror
-----------------------------------------------

Kaj torej narediti?  Poiskati slabe bloke najprej na enem in nato na 
drugem disku in
ponovno vzpostaviti polje?

Zamenjava diskov pomeni popolno reinstalacijo in konfiguracijo, čemur bi 
se rad izognil.

LP., Jurij