vgimport problem

[Victor Makukha]

I have the following configuration: RP8420 and RP7420 connected to the EVA4000. I have B3935DA A.11.16.00 Serviceguard software installed on both nodes. I try to export a VG from one node to another one using "Managing Servicequard Version A.11.16 Eleventh Edition Second Printing B3936-90079":

Source node (RP8420):

Test VG info:
vgdisplay -v /dev/vglock
--- Volume groups ---
VG Name /dev/vglock
VG Write Access read/write
VG Status available
Max LV 255
Cur LV 1
Open LV 1
Max PV 16
Cur PV 1
Act PV 1
Max PE per PV 1016
VGDA 2
PE Size (Mbytes) 4
Total PE 255
Alloc PE 254
Free PE 1
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0
--- Logical volumes ---
LV Name /dev/vglock/lv_lock_test
LV Status available/syncd
LV Size (Mbytes) 1016
Current LE 254
Allocated PE 254
Used PV 1

--- Physical volumes ---
PV Name /dev/dsk/c16t0d1
PV Name /dev/dsk/c17t0d1 Alternate Link
PV Name /dev/dsk/c18t0d1 Alternate Link
PV Name /dev/dsk/c19t0d1 Alternate Link
PV Name /dev/dsk/c20t0d1 Alternate Link
PV Name /dev/dsk/c21t0d1 Alternate Link
PV Name /dev/dsk/c22t0d1 Alternate Link
PV Name /dev/dsk/c23t0d1 Alternate Link
PV Status available
Total PE 255
Free PE 1
Autoswitch On

Then I do:

# umount /FS_Test
# vgchange -a n /dev/vglock
# vgexport -p -s -m /tmp/vglock2.map /dev/vglock

map file contents:

VGID 4a796b2844687a81
1 lv_lock_test

On node 2 (RP7420):

We see EVA4000 as the following:
# ioscan -fnC disk

Class I H/W Path Driver S/W State H/W Type Description

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

disk 0 0/0/0/3/0.5.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t5d0 /dev/rdsk/c0t5d0
disk 1 0/0/0/3/0.6.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c0t6d0 /dev/rdsk/c0t6d0
disk 7 0/0/10/1/0.98.0.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c21t1d2 /dev/rdsk/c21t1d2
disk 8 0/0/10/1/0.98.1.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c23t1d2 /dev/rdsk/c23t1d2
disk 9 0/0/10/1/1.97.0.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c13t1d2 /dev/rdsk/c13t1d2
disk 10 0/0/10/1/1.97.1.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c15t1d2 /dev/rdsk/c15t1d2
disk 2 1/0/0/3/0.6.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c4t6d0 /dev/rdsk/c4t6d0
disk 3 1/0/0/3/1.2.0 sdisk CLAIMED DEVICE _NEC DVD+RW ND-2100AD
/dev/dsk/c5t2d0 /dev/rdsk/c5t2d0
disk 4 1/0/1/1/0/1/1.6.0 sdisk CLAIMED DEVICE HP 73.4GST373453LC
/dev/dsk/c7t6d0 /dev/rdsk/c7t6d0
disk 5 1/0/2/1/0/4/0.0.0 sdisk CLAIMED DEVICE HP LOGICAL VOLUME
/dev/dsk/c8t0d0 /dev/rdsk/c8t0d0
disk 6 1/0/2/1/0/4/0.0.1 sdisk CLAIMED DEVICE HP LOGICAL VOLUME
/dev/dsk/c8t0d1 /dev/rdsk/c8t0d1
disk 11 1/0/4/1/0.98.0.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c17t1d2 /dev/rdsk/c17t1d2
disk 12 1/0/4/1/0.98.1.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c19t1d2 /dev/rdsk/c19t1d2
disk 13 1/0/4/1/1.97.0.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c9t1d2 /dev/rdsk/c9t1d2
disk 14 1/0/4/1/1.97.1.0.0.1.2 sdisk CLAIMED DEVICE HP HSV200
/dev/dsk/c11t1d2 /dev/rdsk/c11t1d2

Then I try to do the following:
# vgimport -s -m /tmp/vglock2.map /dev/vglock
Couldn't access the list of physical volumes for volume group "/dev/vglock".

Ok, we have a warning in the doc: "Note that the disk device names on ftsys10 may be different from their names on ftsys9. You should check to ensure that the physical volume names are correct throughout the cluster." - but it says "correct" not "the same", but I tried the following in accordance with the man on vgimport:

# vgimport -m /tmp/vglock2.map /dev/vglock /dev/dsk/c9t1d2
vgimport: Warning: Volume Group contains "0" PVs, "1" specified. Continuing.
vgimport: Quorum not present, or some physical volume(s) are missing.
Logical Volume is not defined on any physical volume.
"/dev/vglock/lv_lock_test" is missing Physical Volumes.
Warning: A backup of this volume group may not exist on this machine.
Please remember to take a backup using the vgcfgbackup command after activating the volume group.
# vgdisplay -v /dev/vglock
vgdisplay: Volume group not activated.
vgdisplay: Cannot display volume group "/dev/vglock".
# vgextend /dev/vglock /dev/dsk/c11t1d2
vgextend: The volume group "/dev/vglock" is not active. Only an active
volume group can be extended.
# vgchange -a y /dev/vglock
vgchange: Warning: Couldn't attach to the volume group physical volume "/dev/dsk/c9t1d2":
Cross-device link
vgchange: Warning: couldn't query physical volume "/dev/dsk/c9t1d2":
The specified path does not correspond to physical volume attached to
this volume group
vgchange: Warning: couldn't query all of the physical volumes.
vgchange: Couldn't activate volume group "/dev/vglock":
Quorum not present, or some physical volume(s) are missing.

The same situation happens if I try to indicate all 8 devices or if I do the same through the SAM.

If I create the VG on the same EVA LUN from the 2nd node from the scratch - it is successful (so the devices are accessible). Read-write operations are OK. Looks mysterious.

Any ideas?

[Chan 007]

Hi,

1st you need to create the vg using vgcreate & mknod

also just check diskinfo /dev/rdsk/cxtydz

and then do you vgimport

Chan

[Rajesh SB]

Hi,

On node2,
1. Ensure all the disks are visible in CLAIMED status.

2. Create Volume Group and group device file each VGs
# mkdir /dev/
# mknode /dev//group c 64 0x00000n

3. Import the LV information using exported map file from node1.

Regards,
Rajesh

[Vivekanandan Somanathan]

Hi,

Check for the /etc/lvmpvg content for the particular volume group.

- Vivek

[Victor Makukha]

Hello everyone,

Pls. note , I've created everything following the standard rules, i.e. created dir, group file with the same minor number and so on...

Regards,

Victor

[Ninad_1]

Hi Victor,

Is the VG activated in server 1 when you are trying to import on server2 ?
If yes then first deactivate it and then try vgimport on node2 - I think this is what you have done - but still it complains saying that it is not able to find any disk with the vgid mentioned in the mapfile.
So can you try one thing - try exporting the VG from node1 and then try importing it on node 2.
Also you have trid giving the PV Name specifically - but how did you determine that this is the correct disk - the same disk containing vglock ?

Regards,
Ninad

[Chan 007]

Hi,

Do diskinfo /dev/rdsk/c9t1d2, what is the output? If you don't have the output, then present the LUN to this server and follow your procedure

Chan

[Enrico P.]

Hi,

Are you sure that the disk int the firt an second node is the same?

Enrico

[Enrico P.]

Hi,

Are you sure that the disk int the first an second node is the same?

Enrico

[Victor Makukha]

On node 1:

# diskinfo /dev/rdsk/c16t0d1
SCSI describe of /dev/rdsk/c16t0d1:
vendor: HP
product id: HSV200
type: direct access
size: 1048576 Kbytes
bytes per sector: 512

On node 2:

# diskinfo /dev/rdsk/c9t1d2
SCSI describe of /dev/rdsk/c9t1d2:
vendor: HP
product id: HSV200
type: direct access
size: 1048576 Kbytes
bytes per sector: 512

Looks the same but scantily. Is there any way to get more detailed information on these devices (they are connected via fibre channel adapters)?

[Chan 007]

Hi,

while doing vgimport

vi /tmp/vg_lock.txt & add all disks into that one file.

cat /tmp/vg_lock.txt
/dev/dsk/c9txdy
/dev/dsk/c9txdz
/dev/dsk/c11txdz

then

vgimport -v -m -f /dev/vg_lock

Hope this helps

Chan

[Ninad_1]

Even if you mention all paths while importing the VG I doubt it will help.

Did you try exporting the VG from the node1 and then importing on node2 ?

The root issue seems that the disk /dev/dsk/c9t1d2 is not having the same vgid somehow. But I think it is having a valid vg info as it did not give error while importing using that device name.
Also when you try importing a very strange message saying that there are 0 PVs in the VG on disk c9t1d2.
To confirm that there is no issue of node2 not able to read or take control of disk - try exporting the VG from node 1 and then import in node2

Regards,
Ninad

[Sandman!]

Hi Victor,

It's possible that the device files for the shared disks are missing on node 2 (RP7420). Try recreating them first and then importing the VG as follows:

# insf -e -C disk
# vgimport -s -m /tmp/vglock2.map /dev/vglock

[Denver Osborn]

What's up with vgimport warning: "contains 0 PVs, 1 specified"... maybe something is wrong with the PV header.

Can you check the Pv for the VGID...

echo 0x2010?2X|adb /dev/dsk/c16t0d1 |expand|tr -d " "|sed "s/2010:/VGID /"

If it doesn't show the VGID as seen in the map file, then vgcfgrestore the header to the PV.

on source node...

vgcfgrestore -n /dev/vglock /dev/rdsk/c16t0d1

then on dest node...

ioscan -fn
insf -e
mkdir /dev/vglock
mknod /dev/vglock/group c 64 0xNN0000 <-- where NN is unique hex number between 0 and maxvgs


vgimport -v -s -m /tmp/vglock.map /dev/vglock


Hope this helps,
-denver

[Victor Makukha]

Hi everyone,

Thank all of You for the hints, the problem was 2 different LUNs of the same size presented to these nodes.

Regards,

Victor.

[Victor Makukha]

OK, I'll close this thread then...