Already recorded in etc lvmtab file

[chindi]

Hi ,
We have BC clone server where we mount clone luns.
We are getting disk is already recorded in etc lvmtab error.
We have tried recreating the same , then we tried vgreduce -f vgname , then cd /dev and rm -f vgname , then rebooted server.

But still we are not able to vgimport clone luns

Storage is vnx 5500 , OS hpux 11.31 , No multipathing sftware.
Part of SG node , but other node does not contain clone luns mapped to it .

[chindi · Accepted solution]

vgimport: The physical volume "/dev/disk/disk301" is already recorded in
the "/etc/lvmtab" file.

We rebooted twice , recreated lvmtab again.

Issue resolved

[Torsten.]

I'm not sure if I understand your problem, but you cannot import a clone of a VG that is already mounted because of the same VGID. Change the VGID and it can be imported/activated.

[chindi]

Hi Torsten,

We have a db server of which we take mirror clone and mount it on a backup server.

We were facing issues in mounting clones on BC server .

Due to many invalid disks , do we have any single command to identify problematic vg in /etc/lvmtab file .

[Torsten.]

So maybe the clone is always seen as a new disk with the same VGID. Do you export the VG after backup (vgexport)? Do you change the VGID of the clone (vgchgid)?

[chindi]

Yes we do vgexport daily.

We are not using vgchgid .

Do we have any single command to identify problematic vg in /etc/lvmtab file .

[Torsten.]

>> Do we have any single command to identify problematic vg in /etc/lvmtab file .

I think there is nothing wrong or corrupt until you try to import a VG with the same VGID again. Change VGID on the clone and it should work.

[chindi]

Our problem is resolved after rebooting host and recreating lvmtab.

Just wanted to know how to identify that particular "vg" posing problems .

[Torsten.]

You may be have another problem when presenting a LUN from an array, unpresent and present again or present another LUN with the same number like a LUN before. The server will detect a "known" WWN and deny access until you do a

scsimgr -f replace_wwid ...

or just reboot the server.

Check your scripts again, please!

Check the syslog for

... Run 'scsimgr replace_wwid' ...

messages.

Depending on what you really want to do consider to

- clone the LUN

- change VGID

- replace/validate WWN

- backup

- vgexport the clone

[chindi]

Hi Torsten ,

We do have scsimgr replace_wwid messages in our syslog.

As per your suggestion will look at scsimgr -f replace_wwid ...

As of now rebooting that server was the option that we had opted for.

We do have problem when presenting a LUN from an array, unpresent and present again or present another LUN with the same number like a LUN before.

Its been troubling us for many days especially with EMC VNX5500 & 11.31 .