Kernel Oops!!!

[Άγγελος Οικονομόπουλος]

On Wednesday 11 December 2002 20:29, Pavlos Parissis wrote:
> Hello all,
> Ta provlimata sinegizontai:(
> Shmera paw sto server pou anefera to provlima me to file system kai to
> vriskw kolimeno!
> Kanw hardware reset kai efaga 2 wres sta check tou file system.
> Akoma na katalavw giati inode me provlima ta opoia ta kanw clear meta
> ksana kanw fsck kai pali ta
> idia.
> Na anaferw pws shmera otan ekana fsck sto megalo partitions otan teleiwse
> to elegxo twn files
> ksana ekana mono to check,fsck restarting na thimame kala htan to
> message!! Ekana check gia bad blocks kai sta /dev/ataraid/d0pN devices
> alla kai stous diskous /dev/hde kai /dev/hdg.
> Den peira kanena error.
> Ara lew kati den ginete swsta edw , den mporei na exw sinexeia provlima
> ta inodes kai oi diskoi na mhn
> exoun provlima.

a8we :) den eixes pote reiserfs me ena _sugkekrimeno_ via, e?

> Psaxnontas to logs eida afto:
>  Dec 10 20:19:25 mainserver kernel: Unable to handle
>  kernel NULL pointer dereference at virtual address 0000000c
>  Dec 10 20:19:25 mainserver kernel:  printing eip:
>  Dec 10 20:19:25 mainserver kernel: c01297d0
>  Dec 10 20:19:25 mainserver kernel: *pde = 00000000
>  Dec 10 20:19:25 mainserver kernel: Oops: 0000
>  Dec 10 20:19:25 mainserver kernel: CPU:    0
>  Dec 10 20:19:25 mainserver kernel: EIP:   0010:[<c01297d0>]    Not
> tainted Dec 10 20:19:25 mainserver kernel: EFLAGS: 00010046
>  Dec 10 20:19:25 mainserver kernel: eax: 01600000 ebx: 40000000   ecx:
> 00000000   edx: 00000000
>  Dec 10 20:19:25 mainserver kernel: esi: 00000000
>  edi: 00000206   ebp: d1a106c0   esp: df7f1e64
>  Dec 10 20:19:25 mainserver kernel: ds: 0018   es: 0018   ss: 0018
>  Dec 10 20:19:25 mainserver kernel: Process kjournald pid: 131,
[snip]
> !!!
> Thn wra 20:19 eixa kinisei data apo samba simfona me ta logs.

dhladh to mhxanhma etrexe 2 meres me oopsed kernel (kala, ta logs den ta 
koitas pote? mono otan uparxoun problhmata?). kai liga problhmata eides sto 
filesystem. mia kalh idea 8a htan na trabh3eis ena mkfs.

> Lew hrthe h wra na valw ton 2.4.20 kai ksekinaw.
> Paw na kanw compile ton kernel 2.4.20,eixa 2.4.19 kai pernw segmentation
> fault
> otan ekana make dep!!!
> Ksana kanw make mrproper make xconfig kai pernw segmentation fault otan
> dinw save!!!!
> ksana pali kai to pc kane reset mono tou env ekana compile!!!!!!!!
> Etsi opws hmoun travaw mia sto grafeio kai me akouse olh h etairia kai
> eginA
> rezili se ksenouns anthropous:(

kalutera na 3eroun me poion exoun na kanoun :-)

> Kanei reset tp pc eftixos den xreiaztike na kanw fsck pali kai paw

ti ennoeis? 9 grammes pio panw egrapses oti to mhxanhma ekane reset mono 
tou. tre3e fsck *etsi kai alliws*, to log replay tou ext3 den se 
diasfalizei apo phruna pou den leitourgouse swsta otan krasare.

> console kai kanw compile xwris kanena provlima!
> An mexri thn paraskevh den exei kolisei profanos kati tha eixe o 2.4.19,
> an kanei reset h exei kolisei omws tha parv to DDR chipaki kai tha to
> faw. H na parw kai thn mboard(asus a7v333)???

kaneis den 3erei, giati den exeis *kan* apokwdikopoihsei to oops (man 
ksymoops). mporei na einai kapoios controller, na ta exei pai3ei to chipset 
h na ftaiei h mnhmh. 8a htan polu pio eukolo na brei kaneis ti ftaiei an 
mporouses na dwseis to trace apokwdikopoiwntas to oops. etsi, an doume oti 
o pointer pernaei null check epeidh einai 0xc, mporeis na eisai sxedon 
bebaios oti h mnhmh(to bus? o controller?) biazetai na sunanthsei th megalh 
management unit ston ourano.

>
> Genikh erwthsh, otan pernoume tosa segmentation faults kai Kernel Oops
> poios 'h poioi mporei na ftaine?RAM 'h CPU 'h HDD ti apo olla?

isws kapoio ap' ola, isws kanena. trexeis memtest86, ta diagnwstika tou 
kataskeuasth gia to disko sou, bgazeis ena-ena ta e3arthmata mexri na breis 
ti ftaiei klp.

>
> Egw eimai sxedon sigouros oti kati den pai kala me to hardware.

einai giati eisai polu aisiodo3os. egw eimai sxedon apoluta sigouros oti 
kati ftaiei me to hardware, alla den nomizw oti se mhxanhma pou exei 
hostname "mainserver" prepei na katseis na to psa3eis.

> Pavlos

-- 
Don't diddle code to make it faster - find a better algorithm.
            - The Elements of Programming Style (Kernighan & Plaugher)