Lvextend: could not re-synchonize logical volume

[AwadheshPandey]

Thanks Again,

lvsysnc o/p is same as above.

Regards,

Awadhesh

[A. Clay Stephenson]

At this point, you are almost certainly going to have to lvremove the LVOL and then recreate it and build a filesystem on it and restore from backup. ALWAYS run lvdisplay -v BEFORE doing any LVM operation. Before removing any disk for replacement, it is essential that all the extents on the remaining mirror be "current". Replacing a disk before doing this made recovery impossible.

[Nguyen Anh Tien]

Hi Awadhesh
To determine the cause correctly pls give me output of below commands:
#vgdisplay -v vgname --- vg contains you lv
#ioscan -fnC disk

It is suspected that PV (phisical volume- disks or lUNS) has problem!

[Mridul Shrivastava]

Awadesh,

U pls do lvreduce and remove that pv which is showing some stale extents.
Then try dd to check the disk whether it has some bad sectors.
Then extend the lv again to that pv, i hope this will sovle ur issue. Ur data will be intact.

[Devender Khatana · Accepted solution]

Hi,

The problem here is in the primary copy which is not allowing data to be read from the disk for mirroring onto the second. The solution will be to remove LVOL and recreate it on the second disk which at present have STALE PEs. Doing so will allow you to use the LVOL from one copy for time being and as soon as the faulty disk is replaced extend the mirror onto it.

Also you can have a read test on the faulty disk by

#dd if=/dev/rdsk/cxtydz of=/dev/null bs=1024K

Where /dev/rdsk/vxtydz is the disk having first copy of this LVOL not the one you are extending to.

HTH,
Devender

[singvey]

first , you should check your source volume
fsck -F vxfs|hfs -o full,nolog /xxxxx

second , you can also try this : lvcreate -l 1 -n test vg0x; lvextend -m 1 /dev/vg0x/xvolxxxx pv-path ; lvextend -L bigger-size /dev/vg0x/xvolxxxx; lvsync /dev/vg0x/xvolxxxx

[singvey]

first, you are sure that the second disk size is bigger than the source disk


you should check your source volume
fsck -F vxfs|hfs -o full,nolog /xxxxx

you can also try this : lvcreate -l 1 -n test vg0x; lvextend -m 1 /dev/vg0x/xvolxxxx pv-path ; lvextend -L bigger-size /dev/vg0x/xvolxxxx; lvsync /dev/vg0x/xvolxxxx

[AwadheshPandey]

Sorry Arun,

I accedently assign you 0 points, wating for ur post before closing this thread.

Regards,

Awadhesh

[Arunvijai_4]

Hi Awadhesh,

Here is my post :-)

-Arun

[AwadheshPandey]

Actually I want another solution rather than destroy that lvol. dd worked fine on both the disks, but there was 2 bad block entries in BBDIR of primary disk, thats why 2 PEs are in stale position, Devender,s reply very close to this problem. Thanks to All of you

Regards,

Awadhesh