vgimport problems

[Kim Tran_4]

My system (IA box rx2600, OS: 11.23) connects to an xp disk array. One of the disks (/dev/dsk/c4t6d1, which is also a boot disk) contains a file that I would like to copy to my boot disk. I try to import that disk but failed:
# mkdir /dev/vgtemp
# mknod /dev/vgtemp/group c 64 0x050000
# vgimport /dev/vgtemp /dev/dsk/c4t6d1
Physical Volume "/dev/dsk/c4t6d1" contains no LVM information
Please advise. Thanks.

[saju_2]

Hi kim

First you have to create pvinfo in the disk by pvcreate /dev/rdsk/c4t6d1. If this disk is used by some other server or already some LVM header is there , pvcreate command will fail. But u told that it is a boot disk, then how it will get imported to the system..

Regards
CS

[Madhusudhan_3]

Hi Kim,

1. Is the disk /dev/dsk/c4t6d1 part of any volume group in this system or any other system?

2. Is this only the disk in that volume group or any other disks also exists?

3. Can you try "pvdisplay /dev/rdsk/c4t6d1"
What does it shows?

regards,
madhu

[Kim Tran_4]

Thank you for all your help. The problem is caused by the device file name. 11.23 on IA box uses section 2 to store lvm information. So the device file name should be /dev/dsk/c4t6d1s2. For PA boxes, no section is required.

[vinod_25]

Hi kim

i have read in one of the tech forum that a similar problem has occoured and they have done the following... its FYI

A tusc trace indicated that a READ of the lvmrecord for the VolumeGroup failed. The READ ended with "errno 22, EINVAL", which means "No such device or address". When the vgimport(1M) was run a second time, then
this device would pass fine and the next (newly created) LUN would fail.

(Performing the vgimport(1M) twice as many times as the number of newlyadded LUNs would force the the vgimport(1M) to complete successfully.

Also, if a diskinfo(1M) was issued against the device first, the vgimport(1M) would succeed the first time.)

This strange behavior was resolved by applying the following SCSI patch:

[PHKL_30510/PACHRDME/English]
s700_800 11.11 SCSI IO Cumulative Patch

This patch has been superseded by:

[PHKL_32090/PACHRDME/English]
s700_800 11.11 SCSI IO Cumulative Patch

Note: Please apply this patch and any required dependencies. This patch, as with any patch, may be superseded. Please check for the latest patches at HP's IT Resource Center (ITRC) at the following web site:

http://www.itrc.hp.com


These patches cause certain SCSI commands to retry in cases where the device may have pending check conditions. (Certain SCSI commands may answer with a check condition.)

Good Luck!!!

Vinod K

[Deepak.R · Accepted solution]

Hi,

Kim has got himself out of the trouble time , but I will still try to spread some light on the issue

rx2600 is Itanium based system with HP-Ux 11iv2 (B.11.23) . Here, as Tim mentioned , is a slight difference in the disk device nomenclarute if the disk is boot disk.

An Itanium boot disk having device file name c4t6d1 would have 3 partitons in it as follows

c4t6d1 ( disk device file name )
idisk utility in hpux 11.23 would be using this device file name .
1 - c4t6d1s1 ( EFI partition - FAT )
2 - c4t6d1s2 ( HP-Ux partition )
3 - c4t6d1s3 ( HPSP - HP Service ptn - FAT)

Here, first partition c4t6d1s1 is Eextensible Firmware Interface (EFI) partition is an extension of system firware and contains EFI boot utilities.

Second partition c4t6d1s2 is the HPUx partions and LVM would be using this device file name for all operations . Yoy have got to use this device file name when u try importing VG .

Third one c4t6d1s3 is an optional service partiton , I am very sure about the contents.

Use the command idisk -p /dev/rdsk/c4t6d1 to get the partiton information on this disk.

Pls see the attached disk layout diagram for more clarity

thanks
deepak