TF/CLONE with NoCopy option - Unable to view data on target volume after update on source volume

Ok..Then what would be the difference between TF/CLONE created with NOCOPY and COPY options.

I was able to verify with a clone target created using COPY option. ( had to activate/establish for every refresh to target volume )

I was trying to understand the NOCOPY option when only the new writes/update of existing data would be copied to target volume after we establish the clone session...

Am I missing some thing here !??

Thank you. I see the point that it NOCOPY is something similar to COW snapshots ( TF/SNAP or TF/VPSNAP )

But in my case, when I use clone with NOCOPY, how do I make sure the target devices have the new writes that is different from when I'm using COPY option....or is it that we should  limit the use of CLONE for COPY and for actual NOCOPY benefit we should go with either TF/SNAP or TF/VPSNAP ??

The benefits of a nocopy clone are similar to a snap in that you do not require the full capacity of the source volume to be provisioned or synchronized to the target volume. Obviously these benefits are only fully realized when using virtual provisioning, and in my opinion you will see more benefits of using VP Snaps (a newer version of the nocopy clone).

The create action of TimeFinder Clone creates the point in time on the target volume. If the source volume is updated and you want the target to reflect a later point in time of the source volume you will have to perform a recreate/activate again. Just like any clone/snap of a source, the target is a static point in time.

Any clone target that is activated will copy the writes to the source device to the target device.  If we didn't it wouldn't be a point in time copy.

Regardless of the copy method (-copy or -nocopy) a clone is a single point in time. If you want the target to capture new writes to the source you have to issue recreate/activate. This is the same whether the clone was created with -copy or -nocopy.

Whether you want the target device to be a full clone copy or behave like a snap will depend on your requirements.

Little bit confused now..

if we have to issue establish/activate regardless of copy method ( -copy or -nocopy) .. what is the typical use case of -nocopy 

Even with -copy ( and differential flag set), we have to establish/activate to have the new writes on source copy to target.

Thank you.. I think I'm understanding now..:-) ... In this case how is Clone with -NOCOPY different than TF/SNAP !?

Is this just a COW implementation in CLONE !??

Indeed they are quite similar. There are subtle differences. SAVE devices are required to use SNAP for example. If you have a thin environment, clone would likely be the first choice.

Thanks Jasonc. One more query..How would I know/establish ( any SYMCLI command !? )  that NOCOPY is indeed not copying the actual data but merely pointers and whenever there is a write it actually writes to a target thin pool device(s)..

How should I interpret the following three results ! What does the Copy% indicates !?

Result:1

# symclone -g srini_clonedg_276 query

Device Group (DG) Name: srini_clonedg_276

DG's Type             : REGULAR

DG's Symmetrix ID     : 000492600276

Source Device                   Target Device            State     Copy

--------------------------------- ---------------------------- ------------ ----

               Protected Modified                Modified

Logical   Sym  Tracks    Tracks   Logical   Sym  Tracks   CGDP SRC <=> TGT  (%)

--------------------------------- ---------------------------- ------------ ----

DEV001    1956     32448        0 TGT001    1957        0 .X.. CopyOnWrite    0

Total           -------- --------                --------

  Track(s)         32448        0                       0

  MB(s)           2028.0      0.0                     0.0

Result:2

#symclone -g srini_clonedg_276 query

Device Group (DG) Name: srini_clonedg_276

DG's Type             : REGULAR

DG's Symmetrix ID     : 000492600276

Source Device                   Target Device            State     Copy

--------------------------------- ---------------------------- ------------ ----

               Protected Modified                Modified

Logical   Sym  Tracks    Tracks   Logical   Sym  Tracks   CGDP SRC <=> TGT  (%)

--------------------------------- ---------------------------- ------------ ----

DEV001    1956     18512        0 TGT001    1957        0 .X.. CopyOnWrite   43

Total           -------- --------                --------

  Track(s)         18512        0                       0

  MB(s)           1157.0      0.0                     0.0

Result:3

symclone query -g srini_clonedg_276

Device Group (DG) Name: srini_clonedg_276

DG's Type             : REGULAR

DG's Symmetrix ID     : 000492600276

         Source Device                   Target Device            State     Copy

--------------------------------- ---------------------------- ------------ ----

               Protected Modified                Modified

Logical   Sym  Tracks    Tracks   Logical   Sym  Tracks   CGDP SRC <=> TGT  (%)

--------------------------------- ---------------------------- ------------ ----

DEV001    1956         0        0 TGT001    1957        0 XXX. Copied       100

Total           -------- --------                --------

  Track(s)             0        0                       0

  MB(s)              0.0      0.0                     0.0

A symclone query would show a status of  'CopyOnWrite'.

These 3 queries are not from the same clone session. The Flag settings on the last one indicate this is not a CopyOnWrite clone. The copy % just represents the percentage of source device tracks that have been copied to the target, either by the clone performing a copy operation or by updates to the source in a CopyOnWrite clone.

A couple things to note.

The copy % reports on tracks that have had protection cleared.  When nocopy sessions are activated, protection will be cleared for tracks on thin source extents that are unallocated.  So it may appear that a copy is taking place when all that is happening is protection is being cleared.  We get questions about this in VP Snap where it looks like copy is taking place when no source updates have been made.

VP Snap behaves in Copy_on_write mode and also provides the ability to do recreates and restores, which a Clone -nocopy session cannot do.  So VP Snap may be a better option.  It also uses Thin devices and does not require VDEV/SAVE devs.