Dotaz: Problem s SCSI diskem

[Jan Včelák]

Dneska jsem do serveru pridal novy SCSI disk. Je to Maxtor 74 GB.

Na nom vytvoril partition a naformatoval jako ext3.

Jenze kdyz na nej zacnu neco ukladat, kazdou chvilku to za konzoli zacne hazet hlasky, jejichz pricinu si nemuzu vysvetlit. Soucasne take klesa vykon systemu, pochopitelne.

Tady je jeste cast kern.log - ty same hlasky co na konzili.

Dec 26 19:23:17 zeus kernel: scsi0:0:4:0: Attempting to queue an ABORT message
Dec 26 19:23:17 zeus kernel: CDB: 0x2a 0x0 0x3 0x56 0x66 0xf7 0x0 0x0 0x18 0x0
Dec 26 19:23:17 zeus kernel: scsi0: At time of recovery, card was not paused
Dec 26 19:23:17 zeus kernel: >>>>>>>>>>>>>>>>>> Dump Card State Begins <<<<<<<<<<<<<<<<<
Dec 26 19:23:17 zeus kernel: scsi0: Dumping Card State while idle, at SEQADDR 0x8
Dec 26 19:23:17 zeus kernel: Card was paused
Dec 26 19:23:17 zeus kernel: ACCUM = 0x0, SINDEX = 0x8, DINDEX = 0xe4, ARG_2 = 0x0
Dec 26 19:23:17 zeus kernel: HCNT = 0x0 SCBPTR = 0x6
Dec 26 19:23:17 zeus kernel: SCSIPHASE[0x0] SCSISIGI[0x0] ERROR[0x0] SCSIBUSL[0x0] 
Dec 26 19:23:17 zeus kernel: LASTPHASE[0x1]:(P_BUSFREE) SCSISEQ[0x12]:(ENAUTOATNP|ENRSELI) 
Dec 26 19:23:17 zeus kernel: SBLKCTL[0x6]:(SELWIDE|ENAB20) SCSIRATE[0x0] SEQCTL[0x10]:(FASTMODE) 
Dec 26 19:23:17 zeus kernel: SEQ_FLAGS[0xc0]:(NO_CDB_SENT|NOT_IDENTIFIED) SSTAT0[0x0] 
Dec 26 19:23:17 zeus kernel: SSTAT1[0x8]:(BUSFREE) SSTAT2[0x0] SSTAT3[0x0] SIMODE0[0x8]:(ENSWRAP) 
Dec 26 19:23:17 zeus kernel: SIMODE1[0xa4]:(ENSCSIPERR|ENSCSIRST|ENSELTIMO) SXFRCTL0[0x80]:(DFON) 
Dec 26 19:23:17 zeus kernel: DFCNTRL[0x0] DFSTATUS[0x89]:(FIFOEMP|HDONE|PRELOAD_AVAIL) 
Dec 26 19:23:17 zeus kernel: STACK: 0xe1 0x163 0x109 0x3
Dec 26 19:23:17 zeus kernel: SCB count = 20
Dec 26 19:23:17 zeus kernel: Kernel NEXTQSCB = 5
Dec 26 19:23:17 zeus kernel: Card NEXTQSCB = 5
Dec 26 19:23:17 zeus kernel: QINFIFO entries: 
Dec 26 19:23:17 zeus kernel: Waiting Queue entries: 
Dec 26 19:23:17 zeus kernel: Disconnected Queue entries: 9:1 
Dec 26 19:23:17 zeus kernel: QOUTFIFO entries: 
Dec 26 19:23:17 zeus kernel: Sequencer Free SCB List: 6 0 3 11 5 2 8 7 15 12 1 13 4 14 10 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 
Dec 26 19:23:17 zeus kernel: Sequencer SCB Info: 
Dec 26 19:23:17 zeus kernel:   0 SCB_CONTROL[0xe0]:(TAG_ENB|DISCENB|TARGET_SCB) SCB_SCSIID[0x47] 
Dec 26 19:23:17 zeus kernel: SCB_LUN[0x0] SCB_TAG[0xff] 
Dec 26 19:23:17 zeus kernel:   1 SCB_CONTROL[0xe0]:(TAG_ENB|DISCENB|TARGET_SCB) SCB_SCSIID[0x47] 

... v kus vynechavam  ...

Dec 26 19:23:17 zeus kernel:  29 SCB_CONTROL[0x0] SCB_SCSIID[0xff]:(TWIN_CHNLB|OID|TWIN_TID) 
Dec 26 19:23:17 zeus kernel: SCB_LUN[0xff]:(SCB_XFERLEN_ODD|LID) SCB_TAG[0xff] 
Dec 26 19:23:17 zeus kernel:  30 SCB_CONTROL[0x0] SCB_SCSIID[0xff]:(TWIN_CHNLB|OID|TWIN_TID) 
Dec 26 19:23:17 zeus kernel: SCB_LUN[0xff]:(SCB_XFERLEN_ODD|LID) SCB_TAG[0xff] 
Dec 26 19:23:17 zeus kernel:  31 SCB_CONTROL[0x0] SCB_SCSIID[0xff]:(TWIN_CHNLB|OID|TWIN_TID) 
Dec 26 19:23:17 zeus kernel: SCB_LUN[0xff]:(SCB_XFERLEN_ODD|LID) SCB_TAG[0xff] 
Dec 26 19:23:17 zeus kernel: Pending list: 
Dec 26 19:23:17 zeus kernel:   1 SCB_CONTROL[0x60]:(TAG_ENB|DISCENB) SCB_SCSIID[0x47] SCB_LUN[0x0] 
Dec 26 19:23:17 zeus kernel: Kernel Free SCB list: 8 9 4 3 12 15 10 7 13 2 11 14 0 6 19 18 17 16 
Dec 26 19:23:17 zeus kernel: DevQ(0:0:0): 0 waiting
Dec 26 19:23:17 zeus kernel: DevQ(0:1:0): 0 waiting
Dec 26 19:23:17 zeus kernel: DevQ(0:4:0): 0 waiting
Dec 26 19:23:17 zeus kernel: DevQ(0:6:0): 0 waiting
Dec 26 19:23:17 zeus kernel: 
Dec 26 19:23:17 zeus kernel: <<<<<<<<<<<<<<<<< Dump Card State Ends >>>>>>>>>>>>>>>>>>
Dec 26 19:23:17 zeus kernel: (scsi0:A:4:0): Device is disconnected, re-queuing SCB
Dec 26 19:23:17 zeus kernel: (scsi0:A:4:0): Abort Tag Message Sent
Dec 26 19:23:17 zeus kernel: (scsi0:A:4:0): SCB 1 - Abort Tag Completed.
Dec 26 19:23:17 zeus kernel: Recovery SCB completes
Dec 26 19:23:17 zeus kernel: Recovery code sleeping
Dec 26 19:23:17 zeus kernel: Recovery code awake
Dec 26 19:23:17 zeus kernel: aic7xxx_abort returns 0x2002

Napada vas, cim by to mohlo byt? Nemuze byt chyba na strane disku? Nebo jsem disk spatne pripojil?

Odpovědi

[Jan Včelák]

Tak jeste pokracovani - co se stalo, nez jsem dopsal prispevek.
Plna konzole radek s timto:

scsi0 (4:0): rejecting I/O to offline device

A uplne odumreni systemu. Musel jsem ho natvrdo restartovat.

Co s tim? Diky za jakekoli rady

[Jan Včelák]

Aby toho nebylo malo.

Buffer I/O error on device sdc1, logical block 9461197
lost page write due to I/O error on sdc1

[Peter Golis]

len tak hadam. kabel je ok? zdroj je dostatocny? terminator je napojeny spravne a tam kde ma byt?

[Jan Včelák]

Kabel a terminator je OK urcite. Jsou na nem dalsi 2 disky a zalohovaci mechanika na pasku. Zdroj myslim dostatecny.

[Medved09]

Mate spravne nastavene SCSI ID? neni nahodou zapla terminace primo na disku? nemate nahodou disk pripojeny na konci kabelu a terminaci na nekterem puvodnim zarizeni? Tohle opravdu vypada jako problem SCSI kanalu..

[Jan Včelák]

Vypada to asi takhle:

radic -> kabel
           |
           ---> disk1 (ID 1)
           |
           ---> disk2 (ID 2)
           |
           ---> ten novej problemovej disk (ID 4)
           |
           ---> volne misto
           |
           ---> zalohovaci mechanika (ID 6)
           |
           ---> terminator

Predtimm nebylo na miste problemoveho disku nic.

Na disku jsem nastavil pouze ID, tak aby se netlouklo s ostatnimi. Je to poprve, co pracuju s SCSI diskama, takze je opravdu mozne, ze jsem udelal nejakou ptakovinu.

Rikate, ze to je problem SCSI kanalu, ale je zvlastni, ze BIOS disky pri spousteni rozezna Linux stejne tak. Vypise seznam zarizeni na radici. Problemy se zacnou vyskytovat, az kdyz pripojim svazek a zacnu na nej neco zapisovat.