can't clean phisical volum head

[chyaroslav]

in the first time i have created phisical volume from EMC lun with size 600Gb. when i tried to vgreduce i recieved these warning:
vgextend: Warning: Max_PE_per_PV for the volume group (17724) too small for this PV (19355). It was becouse my vg initially was created from 500Gb luns. Then i have splitted my 600gb lun into the 2 by 300Gb and tried to vgextend again (first 300Gb was on the same phisical path as 600Gb). But i have recieved the same Warning. Another 300Gb (in another path) i have appended successfully. I have tried to pvremove and pvcreate -f with the same result. I think that the header of PV is not cleaned.
part1:echo "0d8200?DY" | adb /dev/dsk/c6t1d0
2008: -756468761 2007 Feb 28 18:47:10
part1:pvremove /dev/rdsk/c6t1d0
The physical volume associated with "/dev/rdsk/c6t1d0" has been removed.
part1:echo "0d8200?DY" | adb /dev/dsk/c6t1d0
2008: -756468761 2007 Feb 28 18:47:10

[Robert-Jan Goossens]

Hi,

Try to remove and re initiate the device files.

# rmsf /dev/dsk/c6t1d0
# rmsf /dev/rdsk/c6t1d0

# insf -e

Regards,
Robert-Jan

[Court Campbell]

You might want to try ioscan again since you removed the lun on the EMC device and created two new lun's. The system may still think that the device file that originally pointed to the 600Gb lun is still 600Gb. You could use the disk info command on the raw device file and see what size it reports.

[chyaroslav]

I have made all of suggested (ioscan, insf) but diskinfo display the same size of lun (600Gb)

[chyaroslav]

To examine the real disk size i have executed special EMC command:
part1:syminq | grep /dev/rdsk/c6t1d0
/dev/rdsk/c6t1d0 M(56) EMC SYMMETRIX 5771 41005F5000 322882560

But diskinfo displays:
part1:diskinfo /dev/rdsk/c6t1d0
SCSI describe of /dev/rdsk/c6t1d0:
vendor: EMC
product id: SYMMETRIX
type: direct access
size: 634233600 Kbytes
bytes per sector: 512

[Denver Osborn]

I've experienced something like this before and HP's answer was "reboot". Here's my thread from that problem...

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=846401

When I started digging into it a bit deeper found it had something to do with vxvm daemons... even though I was using LVM, the vxvm stuff was running. Unless there's a documented fix for it now, reboot is probably the answer if you can get the outage w/out much hassle.

Hope this helps,
-denver