[linux] Chybny disk v RAID - ako najst pricinu resp. chybny sektor

[Matus UHLAR - fantomas]

> Matus UHLAR - fantomas napsal(a):
>> ake boli chyby v logu ked to padlo?

On 13.05.09 09:56, patrik na foral.sk wrote:
> daemon.log
> May  7 13:03:18 gnu-linux mdadm: Fail event detected on md device
> /dev/md0, component device /dev/sdb1
> May  7 13:34:46 gnu-linux mdadm: RebuildStarted event detected on md
> device /dev/md0
> May  7 13:35:31 gnu-linux mdadm: RebuildFinished event detected on md
> device /dev/md0
> May  7 13:35:31 gnu-linux mdadm: SpareActive event detected on md
> device /dev/md0, component device /dev/sdb1
>
> v kern.log som v danom case nenasiel nic podozrive,

nebolo v kern.log nic zaujimave pred danym casom? Mozno doslo k realnemu
problemu predtym, toto uvadza len cas kedy to zistil mdadm (napriklad kedy
sa spustil demon).

> udivil ma ale sposob ukladania logov. kern.log.0 obsahuje len jeden den
> (dnesok) ale kern.1.gz je stary niekolko tyzdnov (24.3.).

cim rotujes logy? toto vyzera na stary debianovsky sposob rotacie cez savelog
z cron jobov (alebo ani tych nie). nakolko kern.log.0 obsahuje dnesok,
zrejme ho niekto zrotoval a nerestartol syslog demona.

v kazdom pripade pohladaj kernelove hlasky o sdb vo vsetkych kernel logoch.

>> (kontrolovat SMART stav ti radili, s tym suhlasim).
>
> zial S.M.A.R.T mi nefunguje.
>
> apt-get -y --force-yes -f install smartmontools
>
> Setting up smartmontools (5.36-8) ...
> Enabling S.M.A.R.T..../dev/sda...unable to fetch IEC (SMART) mode page
> [unsupported field in scsi command]
> (failed)...done.
> Starting S.M.A.R.T. daemon: smartd failed!
> invoke-rc.d: initscript smartmontools, action "start" failed.
> dpkg: error processing smartmontools (--configure):
>  subprocess post-installation script returned error exit status 1
> Errors were encountered while processing:
>  smartmontools
>
> SMART je v BIOS-e povoleny, disk je WD VelociRaptor, cipset je nForce  
> 410 MCP s podporou SATA-II

smartctl by vsak uz nainstalovany byt mal, skus smartctl -a /dev/sdb ako ti
uz radili.

>> nemountuj ho, postvi na neho badblocks vo write mode. Samozrejme ked bude
>> odpojeny z RAIDu. badblocks ma dva write mody - nedestruktivny a
>> destruktivny, druhy trva 4-krat tolko ale je spolahlivejsi.
>
> > badblocks -swv /dev/sdb1
> Checking for bad blocks in read-write mode
> From block 0 to 4192933
> Pass completed, 0 bad blocks found.

podla tohoto je uplne OK. Avsak spolahlivo tu informaciu zistis len ked
otestujes cely disk. Ak mas vsetko v mirrorroch, odpoj sdb odvsadial a skus
badblocks na sdb.

>> Pokial ten disk mozes vymenit za zalozny, urob to, a badblocks skusaj v
>> druhom compe.
>
> Fyzickemu kontaktu so strojom som sa chcel vyhnut, no asi to pre
> istotu vymontujem a testnem utilitou od vyrobcu.

este mozes vyskusat tie veci
-- 
Matus UHLAR - fantomas, uhlar na fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
99 percent of lawyers give the rest a bad name.