EVA 3000 Linux LVM (pvcreate)

[David Spencer_1]

I'm setting up a LUN on my EVA3000 using Linux RedHat3 update 4 with LVM. When I run pvcreate I get an "invalid physical volume" message.

# pvcreate -v -f /dev/sda
pvcreate -- locking logical volume manager
pvcreate -- checking physical volume name "/dev/sda"
pvcreate -- getting physical volume size
pvcreate -- checking maximum physical volume size
pvcreate -- checking partition type
pvcreate -- invalid physical volume "/dev/sda"

What can I do to correct the this message.

[Slawomir Gora]

Hi,
did you try to zeroing the first sector ?
if not try to run command:
dd if=/dev/zero of=/dev/sda bs=512 count=1
and now
pvcreate -f /dev/sda

[David Spencer_1]

Yes I did and that did not work.

# dd if=/dev/zero of=/dev/sda bs=512 count=1
dd: writing `/dev/sda': Input/output error
1+0 records in
0+0 records out

# pvcreate -f /dev/sda
pvcreate -- invalid physical volume "/dev/sda"


I also have tried the sfdisk command.

# sfdisk -l /dev/sda
Disk /dev/sda: 261 cylinders, 255 heads, 63 sectors/track
read: Input/output error
sfdisk: read error on /dev/sda - cannot read sector 0
/dev/sda: unrecognized partition


I also ran a mke2fs to see if it could see the LUN. It appears that it can even though it failed to read block 0.

mke2fs -v /dev/sda mke2fs 1.32 (09-Nov-2002) /dev/sda is entire device, not just one partition!
Proceed anyway? (y,n) y
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
262144 inodes, 524288 blocks
26214 blocks (5.00%) reserved for the super user First data block=0
16 block groups
32768 blocks per group, 32768 fragments per group
16384 inodes per group
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912

Warning: could not read block 0: Attempt to read block from filesystem resulted in short read Writing inode tables: done Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 38 mounts or 180 days, whichever comes first. Use tune2fs -c or -i to override.

[Slawomir Gora]

Hi,

are you sure /dev/sda is correct path for your eva3k disk ?

show me:
1. cat /proc/scsi/scsi
2. cat /proc/partitions
3. dmesg | grep sd

4. Are you using SecurePath ?

[Paul Henderson_2]

This happens on my SUN servers running Linux. The problem is that the SCSI driver module is not loaded by default on RHEL WS3.

Here's what I do:

- boot the system from a "Rescue" floppy or CD or network boot Rescue image and get to the Rescue/SafeMode type of shell (bash)
- /sbin/module-setup

- fdisk /dev/sda
- 'p'rint the partition table

- 'w'rite the partition table

Then reboot your regular network boot image or CD.

Basically, you have to push RHEL to load the scsi driver, so you can label the disk, so that the OS can see it. Go figure...

[David Spencer_1]

My problem is resolved. I changed the failover path for the Qlogic FC adapter to "Path A Failover Only" in the EVA CommandView "Vdisk Active Member Properties.

After making this change I was then able to perform the sfdisk to successfully create the partition, and to set it for Linux LVM (0x8e) type. I then was able to execute pvcreate successfully. Thanks to all for all the assistance and quick responses.

[David Spencer_1]

My problem is resolved. I changed the failover path for the Qlogic FC adapter to "Path A Failover Only" in the EVA CommandView "Vdisk Active Member Properties.

After making this change I was then able to perform the sfdisk to successfully create the partition, and to set it for Linux LVM (0x8e) type. I then was able to execute pvcreate successfully. Thanks to all for all the assistance and quick responses.