warning message when trying to extend a logical volume

[holdahl]

Getting this warning message when trying to extend a logical volume on one of my AIX servers:

0516-622 extendlv: Warning, cannot write lv control block data.

Can something be done to avoid this message?


-holdahl

 

[khalidaaa]

I think your LVCB is corrupted!

Have a look here for more info:

http://www.redbooks.ibm.com/redbooks/pdfs/sg245433.pdf

Since the LVCB holds information, such as the creation date of the logical
volume, information about mirrored copies, and possible mount points in a
journaled file system, if its overwritten, these values become undefined. This
will not be the case if we are using a bigVG since we run off the LVCBs in
VGDA. The following discussion mainly applies to non-big VGs.
Certain LVM commands are required to update the LVCB as part of
completeness of algorithms in LVM. The old LVCB area is first read and
analyzed to see if it is a valid LVCB. If the information is verified as valid LVCB
information, then LVM writes an updated LVCB. If the information is not valid,
then LVM does not perform an update (due to the user data corruption
possibility). Instead, LVM displays a warning, such as:
0516-622 synclvodm: Warning, cannot write lv control block data.
It is important to realize that loss of an LVCB does not prevent manipulation of
a logical volume. Once an LVCB is lost, it is still possible to perform the
following actions on the logical volume:
•extendlv
•mklvcopy
•rmlv
• crfs -d (note this will destroy any information within the LVCB)
The problem with logical volumes with corrupted LVCBs is that they cannot be
completely and reliably imported. We saw in 2.3, “Corruption example 1:
Simple ODM corruption” on page 66, that file system log devices and mount
points were lost when we exported, imported, and ran a mklv. The mklv
caused us to lose our LVCB information, and when an LVCB is corrupt, we
are in the same situation.
If the LVCB is deleted, importvg will still define the logical volume to the
system that is accessing this volume group, and the user can still access the
raw logical volume. However, any JFS information is lost until the LVCB is
re-created. We may also need to create a mount point manually if we are
running importvg on a new system.

Not sure if this will help!

If the integrity of the data that has overwritten the LVCB is not an issue, the
following commands will rebuild the LVCB from the ODM:
echo "AIX LVCB\0" | dd of=/dev/hd# bs=1 count=9
updatelv lv_name vg_name

In this post it says this can be ignored!

http://www.tek-tips.com/viewthread.cfm?qid=755035

Regards,
Khalid

 

[MoreFeo]

Is this a raw logical volume, or does it have a filesystem mounted on it?

If it's a raw one, it could be that the application is using the LV, so the extendlv command cannot update the LVCB.

 

[holdahl]

It is a raw logical volume.
Most likely that the application is using the LV so I get this warning message when trying to extend it.

Thanks for the help.

 

[p5wizard]

When using raw LVs, the application should offset the start of the LV by at least 512 bytes (size of LVCB).

If the app has overwritten the LVCB blocks with real data, you should just leave it as it is, or you risk corrupting the data that the app has put there...



HTH,

p5wizard