vgimport from aEMC TimeFinder Clone?

[Infra team member]

Hi folks,
does anyone has experience with cloning with EMC timefinder clone? The mechanism should be the same when I am using dd or whatever to dump whole disks.
My problem is:
I have an active volumegroup and this should be the whole time in this state.
I clone this vg with EMC timefinder clone to another disk. I will use this clone on another host, but when I try to a vgimport I think I run in trouble, because the exact copy of the disk has also the vg ID and needs to change firstly I think. What are the steps to get a good imported vg on the other host?

Thanks for your help.
Klaus

[Mridul Shrivastava · Accepted solution]

Please refer the man pages of vgchgid.

This is used for changing VGIDs in these scenario's.

[Ninad_1]

Hi,

The first thing to note is when you clone the disks in the VG - am not aware of timefinder (As i understand it is similar to BCVs) - but when you are splitting the clone copies you must have your primary volumes quiesced. which means if they are filesystems they must be unmounted, if its database vols , database must be shutdown.

Next - if you are presenting the disks to another host, its not necessary to change the VG id - unless there has been another VG using the same id on the other host.
In that case you can use
vgchgid /dev/rdsk/disk1 disk2 ...
So mention all the disks in the VG on other host to vgchgid - this will change the VG id on the disks.
man vgchgid.
Then you can create /dev/vgname
the group file in /dev/vgname using mknod and using unique minor number
Then import the VG
vgimport /dev/vgname disk1 disk2 ...
vgchamge -a y /dev/vgname [ Activate the VG ]

Regards,
Ninad

[Infra team member]

many thanks!! vgchgid helps, BUT
it looks like that TimeFinder Clone does not work well. I get this when I try to mount the imported vg:
#mount /dev/vgemctst/lvol1 /emctst
vxfs mount: /dev/vgemctst/lvol1 is corrupted. needs checking

#fsck /dev/rdsk/c3t14d2
fsck: /etc/default/fs is used for determining the file system type
vxfs fsck: not a valid vxfs file system
invalid super-block
vxfs fsck: cannot initialize aggregate
file system check failure, aborting ...

Any ideas?
Thanks in advance,
Klaus

[Ninad_1]

As I said earlier, are you stopping all activity on the volumes when you are splitting the clone volumes to be presented to the other host ? Like shutting down the database or unmounting the filesystems ?

Regards,
Ninad

[Infra team member]

no, every thing was up and running, but I have chosen my ignite vg, which does not has any activities normally (except smthg was started by me ;-)).
Normally should be that I can take online clones whithout shutting down the source (like bring the oracle instance in backup mode ...), so it should work as it is sold.

I will try it with unmounting again, too see what happens.

[Ninad_1]

Yes you are partly correct.
You can clone while everything is running. But when you want to split the clones so as to present them to another host - then you need to make sure that there is no activity - rather shutdown database, unmount filesystems for the time it takes to split the clones , and again you can bring the filesys/database up.

Regards,
Ninad

[Rita C Workman]

You don't mention what version of Solution Enabler your running....so my attached document is for 6.0.2 where you run from the primary side to affect the R2/BCV.
If your running an older version you run from the R2 side without the ~ -rdf option.

Before mounting the bcv (timefinder clone) you must do the fsck.
I do the process in a script...so it first puts my prod oracle database into hotbackup, then it re-associates the bcv disk and syncs it up. When it's done syncing it splits them off, runs the fsck for each and then mounts them up.
Copy of one of them is attached..

I have some old notes about setting up the disk the first time if you want that I can attach it. It's just some old scribble notes...but your welcome to use it. Just let me know.

Rgrds,
Rita

[Rita C Workman]

Here's that other document...please remember, like I said, it's just some old scribble notes for setting up BCV's the first time.

Rita

[Infra team member]

Many thanks to all who helps me so fast.
I have to play around a bit to get more experience with the EMC cloning.

Rita,
your scripts looks like a very good help for me. Thanks for that. I am using also the version V6.0.2.0, so that it is every thing very well.

When I get new results, I will summarize.

Thanks and bye for now.
Klaus

[rariasn]

Hi,

I clone my BCVs disks of the following way:

# vgchange -a n vgbcv (in node target)
# symmir -g group_name establish
# symmir -g group_name split

In node source

# vgexport -p -m /tmp/vgori.map vgori
# scp -p /tmp/vgori.map node_target:/tmp

In node target

# vgexport vgbcv (export destructive=
# mkdir /dev/vgbcv
# mknod /dev/vgbcv/group c 64 nnnnnn

For all disks of vgbcv:

# vgchgid /dev/rdsk/cxtydz /dev/rdsk/... ...

# vgimport -m /tmp/vgori.map /dev/vgbcv /dev/dsk/cxtydz ...

# vgchange -a y vgbcv

For all files systems of vgbcv:

# fsck -y -o full /dev/vgbcv/lvnnn

# mount /dev/vgbcv/lvnnn mount_point

rgs,

ran