RAID sync

[Apollon Koutlidis]

V13 wrote:
> To na einai out-of-sync *einai* en meri curruption sta dedomena (h 
> synepagetai). Kapoia metadata den graftikan eno to leitoyrgiko nomize oti 
> graftikan (efoson den prokeitai gia apotomi diakopi reymatos).
>
> Dose mdadm --examine /dev/sdX kai tha deis kapoy na leei kati toy styl:
>
> Events : 0.141868
> h
> Events : 4368
>
> ean ayto den einai idio se oloys toys diskoys tote kati exei graftei ston enan 
> kai oxi ston allon kai oi diskoi theoroyntai out-of-sync. Katalabaineis oti 
> ayto einai mia pliroforia h opoia brisketai pano sto disko. Ean ayth h 
> pliroforia einai diaforetiki se 2+ diskoys (ara theleis RAID5 me >3 diskoys) 
> tote to array den mporei na ksanaginei assembly. Mazi me ayth th pliroforia, 
> akomi kai an h write cache toy diskoy den kanei reordering sta dedomena poy 
> grafontai, yparxoyn pragmata ta opoia den graftikan sto disko.
>
> To oti to 'anagkases' na to kanei assembly den shmainei oti ola einai OK. Kane 
> ena fsck se otidipote brisketai mesa sto array ean den theleis na exeis  
> dysarestes ekplikseis sto mellon .
>
> <<V13>>
>
> p.s.1 To oti ayto poy prosekses (anagkastika) einai to oti den ginetai 
> assembly to array, den synepagetai pos einai to mono problima poy 
> dimioyrgithike.
> p.s.2 To problima mporei na emfanistei kai me 3 mono diskoys ean oloi exoyn 
> diaforetikh timi ston event counter. Sth praksh mporei na symbei alla 
> kalitera na min to analysoyme
Eipa loipon mia pou ka8omai kai ton mastigwnw shmera sth douleia 
patsarontas ena matso solaria na kanw ena grhgoro peirama anamesa sta 
smpatch & reboot.

Eftia3a loipon 4 arxeiakia tou 1.1GB ( dd if=/dev/zero of=volX bs=16384 
count=65536 ) kai tous trabh3a ena loopback upoulo. Meta ekana fdisk, 
ena primary partition ola ta blocks, type FD (Linux RAID autodetect) 
kai  eftia3a ena RAID5 ws e3hs:

# mdadm --create /dev/md0 --raid-devices=4 --level=5 /dev/loop0 
/dev/loop1 /dev/loop2 /dev/loop3

...kai tou ebala daxtulo na dw an kanei ta idia pou eixa dei sto spiti:

# cat /proc/mdstat
Personalities : [raid6] [raid5] [raid4]
md0 : active raid5 loop3[4] loop2[2] loop1[1] loop0[0]
      3145536 blocks level 5, 64k chunk, algorithm 2 [4/3] [UUU_]
      [=>...................]  recovery =  5.0% (53296/1048512) 
finish=1.2min speed=13324K/sec

Edw loipon blepoume ena arxiko indication oti kanei AKRIBWS THN IDIA 
MALAKIA:

loop3[4]
loop2[2]
loop1[1]
loop0[0]

Parathrhse kaneis oti leipei to index [3]? Gia na ri3oume mia matia kai 
sto superblock:

# mdadm -E /dev/loop0
/dev/loop0:
          Magic : a92b4efc
        Version : 00.90.00
           UUID : 1367748c:880454d8:37959946:e94dce6c (local to host 
jalfrezi)
  Creation Time : Sat Jul 12 16:37:07 2008
     Raid Level : raid5
  Used Dev Size : 1048512 (1024.11 MiB 1073.68 MB)
     Array Size : 3145536 (3.00 GiB 3.22 GB)
   Raid Devices : 4
  Total Devices : 5
Preferred Minor : 0

    Update Time : Sat Jul 12 16:37:07 2008
          State : clean
 Active Devices : 3
Working Devices : 4
 Failed Devices : 1
  Spare Devices : 1
       Checksum : f67d1fe5 - correct
         Events : 0.1

         Layout : left-symmetric
     Chunk Size : 64K

      Number   Major   Minor   RaidDevice State
this     0       7        0        0      active sync   /dev/loop0

   0     0       7        0        0      active sync   /dev/loop0
   1     1       7        1        1      active sync   /dev/loop1
   2     2       7        2        2      active sync   /dev/loop2
   3     3       0        0        3      faulty
   4     4       7        3        4      spare   /dev/loop3

Auth th stigmh de mporw na tou kanw reboot (kai den exw virtualisation 
proxeiro) alla upo tis idies akribws sun8hkes o home server mou 
mperdepse ta diskakia tou. Ektote wstoso einai ola kala kai aga8a (btw 
de nomizw oti suntrexei logos gia fsck, to filesystem to eftia3a telika 
apo thn arxh meta to re-assemble).

</apollon>