vgdisplay: Warning: couldn't query all of the physical volumes. need your help

[apple]

Dear HPUX gurus,
would really appreciate you can help me on this.
1. i wanted to extend one disc to vg05, but when i do pvcreate /dev/rdsk/c32t1d6, it gives me warning, Warning: couldn't query all of the physical volumes.

2.then i extend the disc to vg05 and when i issue the vgdisplay -v vg05, i got this error:
vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c22t0d6":
The specified path does not correspond to physical volume attached to
this volume group
vgdisplay: Warning: couldn't query all of the physical volumes.

3. i have run the vgchange -a y vg05, but to no avail, i have searched inside itrc and find the another thread, but i'm not sure can i use the same solution as per this:
http://forums12.itrc.hp.com/service/forums/questionanswer.do?threadId=1145144&admit=109447627+1207568383981+28353475
by vgchange -a y -q n /dev/vg05
4. is it critical problem? i'm afraid i might lose my existing data.

hope to hear from you. thansk a lot

[apple]

attached is the vg05 output. the physical disc are 2 before, but the when i issue the vgdisplay -v vg05 it reflected the 3 PV as per the output.

[apple]

attached is the vg05 output. the physical disc are 2 before, but the when i issue the vgdisplay -v vg05 it reflected the 3 PV as per the output.

[Jeeshan]

your pv named c22t0d6 is not available.

do you have an entry in /etc/lvmtab of this disk?

check.

BTW, is it in SAN? which SAN you used?

[apple]

hi, when i issue the strings /etc/lvmtab returns me with output as per below:
kedidi:/etc>strings /etc/lvmtab |grep c22t0d6
/dev/dsk/c22t0d6

i'm using hp san storage, command view eva.
hope to hear frm you.

[apple]

1. strings /etc/lvmtab output for vg05
/dev/vg05
/dev/dsk/c22t0d6
/dev/dsk/c32t0d6
/dev/dsk/c26t0d6
/dev/dsk/c34t0d6
/dev/dsk/c24t0d6
/dev/dsk/c24t1d1
/dev/dsk/c34t1d1
/dev/dsk/c26t1d1
/dev/dsk/c32t1d1
2. the new disc that i want to extend is
/dev/dsk/c32t1d6 Active
/dev/dsk/c34t1d6 Active
/dev/dsk/c24t1d6 Active
/dev/dsk/c26t1d6 Active

[Jeeshan]

then you need to lot more query and answers

1. is the disk claimed when you used ioscan -fnC disk?
2. did you cheked there is no dead LUN?
3. did u checked in EVA storage any faults?

[Eric SAUBIGNAC · Accepted solution]

Bonjour Almond,

Maybe situation is not so critical ...

First thing : "good" paths to EVA seem to be c24, c26, c32 and c34. It seems that this missing c22t0d6 is something like a ghost path to the EVA. Maybe you did in the past some reconfiguration of your SAN, right ? Could you post result of "ioscan -f -C ext_bus".

If we take a look at lvmtab, we can see the four good path to luns 6 an 9 + the missing c22t0d6 that correspond to a missing c22 path to LUN 6. More, vgdisplay shows 3 current PV and only 2 active PV : normal, since c22t0d6 does no more exist.

The only LV is 125 LE. We can see that it has 103 PE on lun 6 and 12 PE on lun 9, all of these LE through the 4 good paths. There is a good chance that your LV is not corrupted.

In my understanding c22t0d6 is no more used by vg05. Check with "lvdisplay -v /dev/vg05/lvol1". Do you have any stale PE ? Any LE positionned on a "?" disk instead of /dev/dsk/cxxxx ?

If not, you can safely remove this c22t0d6 with the command "vgreduce -f /dev/vg05".

Tell us

Regards

Eric

[apple]

dear HPUX gurus,
i'm bit a happy. i managed to extend the file system. appreciate you can enlighten me on my queries:
1. is the disk claimed when you used ioscan -fnC disk?

almond -> ioscan -funC disk /dev/dsk/c22t0d6 return me with no value

2. you give me hints, then i tried to mv the /etc/lvmtab and rescan the the vgscan -av
but it gives me this output:
LVMTAB has been created successfully.
If the PV links are configured in the system.
Do the following to resync information on disk.
Vgchange â a y
Lvlnboot â R

after that i'm able to extend the disc, but do we need to do resync or something as per the output?

3. kedidi:/etc>lvextend -l 229 /dev/vg05/lvol1
Logical volume "/dev/vg05/lvol1" has been successfully extended.
vgcfgbackup: /etc/lvmtab is out of date with the running kernel:Kernel indicates 13 disks for "/dev/vg05"; /etc/lvmtab has 12 disks.
Cannot proceed with backup.
why i get this vgcfgbackup? why it cannot proceed with backup?

thanks a lot. my system admin journey is really fulfilling with itrc. thanks a lot.

[Eric SAUBIGNAC]

Almond,

where have you seen such advice "you give me hints, then i tried to mv the /etc/lvmtab and rescan the the vgscan -av" ?

I think it was really a bad idea to do that before "vgreduce -f vg05" :-(

Could you post a new output of "vgdisplay -v vg05" and a new "strings /etc/lvmtab" restricting output to vg05 ?

[apple]

if i didn't do the vgreduce, what is the impact like? will it major?
1. output for strings /etc/lvmtab
/dev/vg05
/dev/dsk/c32t0d6
/dev/dsk/c32t1d1
/dev/dsk/c34t0d6
/dev/dsk/c34t1d1
/dev/dsk/c24t0d6
/dev/dsk/c24t1d1
/dev/dsk/c26t0d6
/dev/dsk/c26t1d1
/dev/dsk/c32t1d6
/dev/dsk/c34t1d6
/dev/dsk/c24t1d6
/dev/dsk/c26t1d6
2. kedidi:/>vgdisplay -v vg05

attached is the ouput. i'm wondering, there are only 3 physical extends. but from the vgdisplay -v output, i can see this:
why Cur PV 4
Act PV 3

what does this mean?

[apple]

if i didn't do the vgreduce, what is the impact like? will it major?
1. output for strings /etc/lvmtab
/dev/vg05
/dev/dsk/c32t0d6
/dev/dsk/c32t1d1
/dev/dsk/c34t0d6
/dev/dsk/c34t1d1
/dev/dsk/c24t0d6
/dev/dsk/c24t1d1
/dev/dsk/c26t0d6
/dev/dsk/c26t1d1
/dev/dsk/c32t1d6
/dev/dsk/c34t1d6
/dev/dsk/c24t1d6
/dev/dsk/c26t1d6
2. kedidi:/>vgdisplay -v vg05

attached is the ouput. i'm wondering, there are only 3 physical extends. but from the vgdisplay -v output, i can see this:
why Cur PV 4
Act PV 3

what does this mean?

[Eric SAUBIGNAC]

Here is the problem :

Cur PV 4
Act PV 3

The /etc/lvmtab is now correctly configured due to "vgscan" command, but the VG itself keeps track of the ghost disk : 4 disks, the "ghost" disk + 2 original disks + the new one.

"vgreduce -f" would have corrected this ... if done before "vgscan". Don't know if it could work well after. Try and tell us ...

Eric

[Jeeshan]

then i guess you need to do vgcfgrestore

[apple]

1. i could rewind the time, how can i apply this the vgreduce command?
vgreduce -f?

2. Vgcfgrestore? if i didn't what is the impact? i want to try, but will it cause our application residing inside the mount point a downtime. normally the backup and restore thing will take time. erm

[Eric SAUBIGNAC]

1) vgreduce -f vg05

2) As far as i understand your problem, "vgcfgrestore" will resolve nothing and may damage your VG : saved configuration differs from real configuration since you add a new disk in vg05

Eric

[apple]

1) vgreduce -f vg05
almond -> if i do vgreduce, will it reduce all my vg. i heard last time, it's dangerous to do the vgreduce. do i need to specify reduce of how much?
2. is it my new disc extension is safe?

thanks a lot

[Sajjad Sahir]

Dear Almond
if u do vgreduce command it will reduce lastly added physical volume
if u want to remove a particular volume
group
vgchange -a n volume group

vgexport --- volume group name
this will remove u volume group

[Eric SAUBIGNAC]

I can't believe it -> "if u do vgreduce command it will reduce lastly added physical volume" ???

vgreduce removes only the specified disks given on the command line !

Now, "-f" option of vgreduce has a special meaning. Here is an extract of "man vgreduce"

----------------------------------
-f vg_name : Force reduction of missing physical volume(s) in a given volume group. vgreduce obtains the name of each physical volume (PV) belonging to the volume group. It then reads the kernel PV structures to work out which PVs are missing. PVs which are missing will be candidates for removal. If all the physical extents on the missing PV are free then it will be removed from the volume group. Otherwise vgreduce will report the physical to logical extent mapping. For missing PVs which have extents in use, you must free up all the extents by using lvreduce or lvremove and re-run vgreduce with the -f option. This option is most commonly used when the vgdisplay command shows "Cur PV" higher than "Act PV" and all of the PVs belonging to the volume group are attached. This option only works on PVs and not on links.
---------------------------------

Good luck
Eric

[apple]

interesting. :) thanks a lot. will assign points in a bit

[apple]

sorry, appreciate your advice. i have vgscan -av, and now i've the new updated /etc/lvmtab with no missing device.
in this case, do i need to vgreduce -f?
thanks a lot

[Eric SAUBIGNAC]

Bonsoir,

Yes. I think that you still have Cur PV = 4 and Act PV = 3 (check with vgdisplay). So "vgchange -f vg05" is adequate in this case.

At the end of the command you will probably get a message saying that you must do a vgscan. It will not be necessary as /etc/lvmtab is OK. Check with "strings /etc/lvmtab"

Regards

Eric