vg --issue-pv not correspnding

[subodhbagade]

Hello Expert....

i having hp-ux 11.11 server and Sap application are running on it..please help me out for

whenever i am putting cmd
1)#vgdisplay

vgdisplay: Warning: couldn't query physical volume (1) "/dev/dsk/c9t0d0":, (2)/dev/dsk/c11t0d0": (3)"/dev/dsk/c9t0d1 (4)"/dev/dsk/c11t0d1
The specified path does not correspond to physical volume attached to this volume group

(2) i have attached o/p of
#strings /etc/lvmtab and
#ioscan -funC disk

all pv are listed in ioscan and all above mention pv are belong to vg /dev/lathstg.

(3) on .# vgcfgrestore -n /dev/lathstg /dev/rdsk/c9t0d0 or other path it showing

vgcfgrestore: IO error on Physical Volume device /dev/rdsk/c9t0d0: # this is for all case.


(4)Vgcfgrestore also not working ......What steps needs to be follow before Vgcfgrestore

Do we need to unmount all lv in vg to deactivate vg ? then

#vgchange â a y /dev/vg00
· Resynchronise logical volume
#vgsync /dev/vg00

(5) how should i proceed in this case .......pppplease help me

Thanks in advance.

Regards,
subodh.

[Rita C Workman]

Since your typing vgdisplay, then I'll presume your vg is already up.

First thing I might try, before I take those steps, is to just type:

vgchange -a y /dev/

Rgrds,
Rita

[LynnG]

Agree with Rita, also have you tried doing either a

pvdisplay -v /dev/dsk/c?t?d?
and/or
diskinfo /dev/rdsk/c?t?d?

[freddy_21]

Hello subodh,

when you run vgdisplay -v /dev/lathstg you can see contain from volumegroup ( ex logical volume)?

please give complete output from vgdisplay /dev/lathstg.

if you can access the disk using c8 and c10, my opinion your link have problem. when i saw lvmtab and your error, you have problem with 1 link ( c9 and c11 ).


Regards
Freddy

[SKR_1]

Check for the Fibre channel cards status with fcmsutil command.

Also check with storage team for zoning issue.

Thanks

SKR

[Fabio Ettore]

Hi,

I also suppose that dd to those LUNs fails with I/O error. Just try

dd if=/dev/rdsk/c9t0t0 of=/dev/null

However it sounds like the problem that Bala C had few days ago:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1285301

Do you have some software to manage the storage EMC Clariion (something like PowerPath)?

As posted from Bala C in his thread check the PowerPath configuration.

HTH.

Best regards,
Fabio

[Ninad_1 · Accepted solution]

Hi,

As the error says - I/O error on /dev/rdsk/c9t0d0 - it seems that there is some problem accessing this and other LUNs.

(I think that you have 2 paths to each LUN - thus c8t0d0, c8t0d1, c9t0d0 and c9t0d1 are via one path and remaining c10.. and c11 are alternate paths)

So essentially you have 2 disks which are giving error - c9t0d0 and c9t0d1


Please try
diskinfo /dev/rdsk/c9t0d0
and similarly for /dev/rdsk/c9t0d1

If the output shows correct size for the disk, then try dd to check that there are no I/O errors for the disks
dd if=/dev/rdsk/c9t0d0 of=/dev/null bs=1024

Similarly for the other disk -/dev/rdsk/c9t0d1

Also try the following commands
echo "0d8208?UU" | /usr/bin/adb /dev/dsk/c8t0d0

echo "0d8208?UU" | /usr/bin/adb /dev/dsk/c9t0d0

echo "0d8208?UU" | /usr/bin/adb /dev/dsk/c9t0d1

This displays the VGID for the VG thats stored on the disks.
The VGID shows for c8 and c9 devices should be the same - which confirms that c9 devices are part of the VG /dev/lathstg


Regards,
Ninad

[OldSchool]

at the end of your other thread, it indicates that you tried to do a vgconfigrestore to add these to vg00, not vglathstg

http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1288485

so, exactly what steps *have* you done to these?

[subodhbagade]

Hello once again..

i had attached the require o/p of

(1)diskinfo,
(2) pvdisplay -v path
(3)ioscan -funC disk
(4)vgdisplay -v /dev/lathstg
VG not listing all lv ,,but bdf showing all lvs of vg /dev/lathstg

can i put this cmd in this case (on running server)
(5)#vgchange -a y lathstg

(6)i had tried the vgcfgrestore on vg
/dev/lathstg not on /dev/vg00.

Please help me ..

Thanks in advance

Subodh.

[OldSchool]

do the ioscan again without the "u"

ioscan -fnC disk

post results