error in syslog file

[nagarjuna reddy]

Hi all,

I am new to this area, pls. clarify my doubts regarding the errors in syslog file.
the errors are like this

vmunix: SCSI: Write error -- dev: b 31 0x011002, errno: 126, resid: 8192

vmunix: LVM: Failed to automatically resync PV 1f011002 error: 5

vmunix: SCSI: Read error -- dev: b 31 0x011002, errno: 126, resid: 1024

pls. reply me, these errors are critical to my database??

[ani007]

Hi,
Looks like a disk problem.To get what disk it is do ll /dev/dsk/* | grep 0x011002.

dd if=/dev/dsk/cxtxdx of=/dev/null bs=4k
Check if it returns any error.
If there is any error then you need to log a H/W case with the vendor.

If this is a SCSI disk the you need to check the following also.

SCSI termination may be loose.
may be SCSI data cable problem.

If this disk is a part of an external storage device , worst case it cud be a problem with the slot of the stoarage device too.

Please don't forget to assign points.
Regards,
Ani

[nagarjuna reddy]

hi thanks for your reply,

i am not familiar with the Sys. Admin commands and all, i am a oracle dba, so pls. let me tell the commands to see the errors etc.

thanks and regards.

[ani007]

#ll /dev/dsk/* | grep 0x011002

send me the o/p

[nagarjuna reddy]

i didn't get anything.

i typed the command

ll /dev/dsk/* | grep 0x11002

is it correct?

[ani007]

you did a mistake

#ll /dev/dsk/* | grep 0x11002 --incorrect
it should be 0x011002


so fire this just copy paste
#ll /dev/dsk/* | grep 0x011002

[nagarjuna reddy]

hi

i got this

brw-r----- 1 bin sys 31 0x011002 Jun 28 2007 /dev/dsk/c1t1d0s2

[ani007]

#dd if=/dev/dsk/c1t1d0s2 of=/dev/null bs=4k

send me the o/p

[nagarjuna reddy]

/dev/dsk/c1t1d0s2: Permission denied
dd: cannot open /dev/dsk/c1t1d0s2

[ani007]

it can take some time to gv u the o/p

[nagarjuna reddy]

i am not the root user here, i think that is why it didn't permit me.

[ani007]

#ioscan -fnC disk|grep c1t1d0s2
#diskinfo /dev/rdsk/c1t1d0s2

[nagarjuna reddy]

1. sh: ioscan: not found.

2. sh: diskinfo: not found

[ani007 · Accepted solution]

May be you dont have the priviledge as you are not root user.what you can do provide the ctd no (c1t1d0s2) to Unix System Administrater, tell then to check this.They will able to issue the command.
If the dd command shows error then contact vendor.

Please assign points remember, this forum is supported by people DONATING their time and energies into solutions for you.you can assign points from 0-10 what ever you want
Regards,
Ani

[nagarjuna reddy]

Thanks Very Much

I really thankful to you, you are right i do not have the root privileges to get the required info. i surely give you the points. and pls. let me know, what are the possible issues i find any thing wrong.

regards,
nagarjuna

[ani007]

Itâ s a h/w error ..Anytime can be happen..so when you will fire ioscan it should be good if it shows CLAIMED, if it shows no_hardware then you need to check.
Next from dd command o/p if it says input o/p error that means you have problem with the disk... then you need to log a case with vendor ...may be they will replace the disk. ALL of these not your task. This should be done by unix admin.as a DBA max. you can give them the CTD no (which u hv now).Even if you donâ t give them the ctd no then also they will able to find it from syslog...So don't waste ur time give me ctd no to them they will take necessary actions.

Regards,
Ani

[Matti_Kurkela]

From the previous commands, we know that "dev: b 31 0x011002" refers to /dev/dsk/c1t1d0s2.

Now, the only reason for a modern HP-UX disk to have "s2" at the end of the device name is when it is a system disk in an Itanium system.

> vmunix: LVM: Failed to automatically resync PV 1f011002 error: 5

This also refers to /dev/dsk/c1t1d0s2: 1f in hexadecimal = 31 in decimal.

Your system disk appears to have been mirrored, and this message indicates one of the mirror halves is not healthy.

If your database is not installed on vg00, the disk failure does not threaten your database directly... but unless the failed disk is replaced, the system will crash if the second half of the mirrored system disk fails too. And a system crash *would* threaten your database along with the rest of the system.

MK

[madhuchakkaravarthy]

hi


brw-r----- 1 bin sys 31 0x011002 Jun 28 2007 /dev/dsk/c1t1d0s2

===============================




execute

# ll /dev/*/group

and identify its particular vg . and do a vgdisplay -v vgname and check for current pv and open pv.

better talk to ur unix admin he might be giving some useful tips to u

regards

MC