pvdisplay -v Displaying ?????

[Dan Taylor_1]

When I run pvdisplay -v on my disk, I get ??????? in some of my Physical extents.

When I try to remove the logical volume, I get an error that the volume doesn't exist. When I go to create a new one, I get an error that the volume already exists. I would venture to guess that there is some confusion with LVM. How can I go through and remove this without performing a mediainit?

Also, this is my main boot drive, so I have to be somewhat delicate, since I administer support remotely.

Thanks in advance to anyone that can answer this.

[Michael Tully]

Hi,

Try to run this and look for the word 'stale' and post the results

# vgdisplay -v /dev/vg00

-Michael

[Dan Taylor_1]

I've since removed the files for the area that was supposed to be there under /dev/vg00/file and raw file.

I have no room left on that drive and have moved the area to another drive. This is in bad form since I like to keep my servers all the same to make maintaining them somewhat easier.

I still need a soluable way to go through and remove those stupid question marks.

Any help would be greatly appreciated.

[Dan Taylor_1]

I've ran the command:
vgdisplay -v vg00

nothing shows any signs of stale.

Any other ideas?

[Animesh Chakraborty]

Hi,
U mean that files under /dev/vg00 is deleted ?crw-r----- 1 root sys 64 0x000000 Aug 7 1999 group .......
Not sure what is the problem still you can try
1. vgcfgrestore -n /dev/vg00 /dev/rdsk/cxtxdx

Regards
Animesh

[Dan Taylor_1]

there was a device file under /dev/vg00 that I removed. This allowed me to create a lv. with the same name.

When I run pvdisplay -v /dev/dsk/c1t2d0, I end up seeing a section where the logical volume used to reside having ????? where the name should be. With the question marks in there, I can't seem to clear them out.

I've been researching through old posts and searching for why this is occuring. nothing seems to make sense as to why this is happening.

Mind you, this isn't the first time I've seen this. Usually, I see it when the Hard drives have been at the site for quit some time and we have to replace one of the drives in a vg. All of a sudden, you see these strange ?????? showing up.

There must be some association in LVM that has become disrupted and it can't figure out what that area is for.

[Animesh Chakraborty]

<>> ....You should remove it using lvremove command not by rm.
Make sure that you have latest LVM path installed.

[Dan Taylor_1]

I tried the lvremove command. It didn't work, so I was stuck doing it the wrong way.

[Andre Da Silva Dias · Accepted solution]

Hi!

We have the same problem with one of servers, and also in vg00.

But only in the mirror disk, which is more strange, and we have no idea how it happened.

We think that the way to solve this is to put the box in Maintenence Mode, remove the mirror copies and remove the disk from vg00. After, pvcreate -B the disk, added it to vg00 and rebuild the mirrors.

Andre

[Dan Taylor_1]

The answer to my own question, for future reference by anyone interested is that if the LVM is out of sync, run the following command to resync the Logical Volumes:

lvchange -ay /dev/vg00/lvol4

This will regenerate /etc/lvmtab and synchronize the data diplayed by pvdisplay.

Again, what I was seeing was:
---Physical Extents ---
PE Status LV LE
0000 current ???? 0000
0001 current ???? 0000

Thanks,
DT

[Sanjay_6]

Hi Dan,

Take a look at the thread below,

http://us-support.external.hp.com/cki/bin/doc.pl/sid=02777e3907b4e5b9e2/screen=ckiDisplayDocument?docId=200000054358939

Hope this helps.

Regds

[Sanjay_6]

Hi Dan,

Here is another thread,

http://us-support.external.hp.com/cki/bin/doc.pl/sid=02777e3907b4e5b9e2/screen=ckiDisplayDocument?docId=200000056114142

Hope this helps.

Regds

[Dan Taylor_1]

Sanjay,
thank you for the links. I've been doing something similar in the interum to finding a better answer. By removing the device files and then running a lvcreate, lvremove, I can free those areas up.

I would prefer to have a command that would re-tie those areas where they are supposed to go, without having to go to all of this extra homework of figureing out what was there before, and performing the above steps. It's a lot of work.

I wonder if this is fixed with any patches?

[Andre Da Silva Dias]

Hi Dan,

I only had the oportunity to see your replies today.
I was able to solve the problem after creating two device files with 0x00000b and 0x00000c under /dev/vg00.... FANTASTIC!

Many thanks,
Andre