cannot deactivate volume group

[Aaron_82]

I am in the process of creating another package for our DB cluster. I created the volume group with lvs and fs. Then I mounted the filesystems just to make sure they would mount. After successfully doing the applyconf I go to deactivate the volume so I can bring the package up but the volume group will not deactivate. Here is the error when trying to deactivate the VG.

vgchange -a n vga13
vgchange: Failed to notify clvm daemon about volume group deactivation - No such
file or directory
Couldn't deactivate volume group "vga13":
No such file or directory

The group file does not conflict with other VGs and I have not exported the vg to the other nodes yet. So I am stumped.

[RAC_1]

Is the VG vga13 there??
Can you do vgdisplay on it?? Can you activate it as vgchange -a e va13??

Anil

[Aaron_82]

Yes it is there. The volume group is activated and cluster aware. The LV's and FS's are there and they can be mounted(unmounted now).

[RAC_1]

Which package this VG belongs to??

First disble the package switching.
Start and stop the package. The package stop should deactivate the group. Post errors if any(package log)

Anil

[Dietmar Konermann]

Have a look at document UMCSGKBRC00013823. It describes some reasons for this problem and how to avoid it.

Best regards...
Dietmar.

[Aaron_82]

It is pretty simple to recreate this volume group so is there a way I can just delete this volume group without deactivating it?

[Sundar_7]

Afraid not. You cannot recreate the volume group without deactivating and removing/exporting it from the system. As long as the VG has an entry in /etc/lvmtab, you cannot recreate it.

one more thought. Make sure your /etc/lvmtab is intact

# mv /etc/lvmtab /etc/lvmtab.OLD

# vgscan -av

[Aaron_82]

I don't want to run vgscan cuz the other deactivated volume groups will be deleted from lvmtab.

[Sundar_7]

oops ! yeah that is right - time for coffee I guess.

[Aaron_82]

Something new now. I mounted the filesystems earlier this morning but now they will not mount. I get this error:

root@hp02b:/dev/vga13
# mount /dev/vga13/a1302 /a1302
/dev/vga13/a1302: No such device or address

# ll /dev/vga13
total 0
brw-r----- 1 root sys 64 0x030001 Jul 16 09:18 a1302
brw-r----- 1 root sys 64 0x030002 Jul 16 09:27 a1303
brw-r----- 1 root sys 64 0x030003 Jul 16 09:27 a1304
brw-r----- 1 root sys 64 0x030004 Jul 16 09:27 a1305
brw-r----- 1 root sys 64 0x030005 Jul 16 09:27 a1306
brw-r----- 1 root sys 64 0x030006 Jul 16 09:27 a1307
brw-r----- 1 root sys 64 0x030007 Jul 16 09:27 a1308
brw-r----- 1 root sys 64 0x030008 Jul 16 09:28 a1309
brw-r----- 1 root sys 64 0x030009 Jul 16 09:30 a1310
brw-r----- 1 root sys 64 0x03000a Jul 16 09:31 a1311
brw-r----- 1 root sys 64 0x03000b Jul 16 09:31 a13arch
brw-r----- 1 root sys 64 0x03000d Jul 16 09:31 a13base
brw-r----- 1 root sys 64 0x03000c Jul 16 09:31 a13exp
brw-r----- 1 root sys 64 0x03000e Jul 16 09:32 a13utl
crw------- 1 root sys 64 0x030000 Aug 4 2003 group
crw-r----- 1 root sys 64 0x030001 Jul 16 09:18 ra1302
crw-r----- 1 root sys 64 0x030002 Jul 16 09:27 ra1303
crw-r----- 1 root sys 64 0x030003 Jul 16 09:27 ra1304
crw-r----- 1 root sys 64 0x030004 Jul 16 09:27 ra1305
crw-r----- 1 root sys 64 0x030005 Jul 16 09:27 ra1306
crw-r----- 1 root sys 64 0x030006 Jul 16 09:27 ra1307
crw-r----- 1 root sys 64 0x030007 Jul 16 09:27 ra1308

As you can see the files are there. Any suggestions

[Sundar_7]

how about running vgscan -av without moving /etc/lvmtab ? - will that work ?

[RAC_1]

vgchange -a e vga13.

Then try mounting it.

Anil

[Sundar_7]

I have seen this problem if you remove the group file and recreate it with some other minor number.

I was not able to mount the LVs but I was able to deactivate the volume group.

[Sundar_7]

Just a thought - did you activate the vg in exclusive mode and halted the cluster without first deactivating the volume group.

I believe volume groups activated in exclusive mode require /usr/lbin/cmlvmd daemon running. I could be wrong here though :-)

[Aaron_82]

I have tried all that was suggested. There was a post I saw earlier today that was exactly like my issue but there resolution was to reboot. Since this is a production box that is not possible. Anyway, I have to have this package done by the end of the day so I am going to vgreduce the pvs and create another volume group for the package. Thanks for all the suggestions and I still have email notification on this thread just in case there is a solution without rebooting.

[Mohanasundaram_1]

Hi Aron,

Did you try doing fuser on the vg13 filesystems to ensure there is nothing waiting because of it?

You had said later, that "no such device" message came when you tried mounting the lvols. This can happen if your VG is not active :)

what is vgdisplay -v vg13 saying now.

Any additional details in syslog.log file for this problem?
Provide the output of the below,
1) cmviewcl
2) vgdisplay -v vg13
3) vgchange -a n vg13

Cheers,
Mohan.

[Massimo Bianchi]

Hi,
the cmlmd is running, right ?

Then, try to re-activate it, even if it is already active. It is used for re-activating the kernel structures.


vgchange -a y vga13

or

vgchange -a e vga13

HTH,
Massimo

[Massimo Bianchi]

Hi,
can you see the disks n the VG ??

[Aaron_82]

I had to abandon this issue to get the package created. I am now using vga13a for the a13 package. There is still one disk in vga13 cuz it won't deactivate but I vgreduce all the others for vga13a.

As requested:
cmviewcl:
CLUSTER STATUS
proddb_cl1 up

NODE STATUS STATE
hp01a up running

PACKAGE STATUS STATE AUTO_RUN NODE
a02 up running enabled hp01a
a01 up running enabled hp01a
a03 up running enabled hp01a

NODE STATUS STATE
hp01b up running

PACKAGE STATUS STATE AUTO_RUN NODE
a10 up running enabled hp01b
a11 up running enabled hp01b
a12 up running enabled hp01b

NODE STATUS STATE
hp02a up running

PACKAGE STATUS STATE AUTO_RUN NODE
a06 up running enabled hp02a
a04 up running enabled hp02a
a07 up running enabled hp02a
a08 up running enabled hp02a

NODE STATUS STATE
hp02b up running

PACKAGE STATUS STATE AUTO_RUN NODE
a16 up running enabled hp02b
a15 up running enabled hp02b
a13 up running disabled hp02b

vgdisplay -v:
--- Volume groups ---
VG Name /dev/vga13
VG Write Access read/write
VG Status available, exclusive
Max LV 255
Cur LV 0
Open LV 0
Max PV 99
Cur PV 1
Act PV 1
Max PE per PV 2157
VGDA 2
PE Size (Mbytes) 4
Total PE 2157
Alloc PE 0
Free PE 2157
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0


--- Physical volumes ---
PV Name /dev/dsk/c36t3d3
PV Status available
Total PE 2157
Free PE 2157
Autoswitch On