Difference between Gold Copy & Target copy

what do you consider "target copy" , just another BCV/Clone hanging off R2 ?

Yes... As like Gold copy... its another BCV/clone to pair with R2.

Another reason for using Gold & Target is... during DR excercises.. as one copy will be busy with DR site... other copy will be used to create consistent copy.

Thanks

Sandeep

gold BCV is still good for situations where corruption occurs on the R1, corruption propagates to R2. You can restore BCV/Clone to R2 and then push it back to R1.

yes, I do understand that.. But Im trying to understand what kind of setup are others using.

Do you still use Gold/Target/R2 at DR site?

we have one gold copy that we clone from R2 every day.

We use R1 - R2 - GBCV.  The R2 is "asynch / consistent".  We split/synch the GBCV on the DR (target) Array every 4 hours.  We conduct our quarterly DR tests on the GBCVs in the DR data center.  We are just now implementing DR directly from the R2s (because our management decided that 4 hour-old data isn't good enough for a real DR), so we're trying to develop/test that now.

Our R1-R2s are always consistent, except every once-a-month the link between Prod - DR data centers goes down for some stupid line error, and then we may have to resynch R2s.  Usually they resume automatically.

We take our backups from the Source Array.  We have an additional set of "BCVBs", R1-BCVB, that we split once a day to mount on the "Backup Server" in the primary data center; then we "netbackup" the BCVBs to an EMC DataDomain (we're phasing out tape).  I'm not involved with that - the Backup Group controls netbackup/DD.  We (Storage Team) controls the SRDF and BCV synch/split scripts.

    Stuart

I may not know the whole story here, but why did you ever have GBCV and TBCV to begin with?

Hi Stuart.. thanks for the info ..

Well to start the story .. we have started using EMC replication from 2005 and then - R2's are never consistent with R1. We used DMX2's, 3's then and lot of network issues, ports down - replication takes time to catch up and many... So.. we used to maintain Golden & target BCVs where consistent copy is taken alternately on each daily..

And as time passed on, we moved to DMX4's but issue was never resolved.. So, we still maintained.

Now, we are on VMAX - and everything looks good.. So, Im looking at suggesting removal of one BCV set..

One question still bothers me.. How is R2 consistent all day ?? We will be in SRDFA/MSC all day... and if there is any network issue - SRDFA will drop and devices will be in TNR mode.. Will the consistency still be maintained??

To create a consistent copy, we do a EMC Timefinder Est/Split (Mainframe commands) ... So, this will ensure me that there is a consistent copy.. but SRDFA/MSC can guarantee a consistent copy still  on R2 due to any outage ??

Im trying to understand how this works for you .. and do you use the same things ? (may be lil different on Distributed)

Thanks

Sandeep

Sandeep:

We are "Consistent" 99% of the time, except about once a month when we have line (telecom) errors.

1.  We run SRDF/A on HP-UX and Linux hosts.

2.  We use a command sequence like this to set them up:

3.  When the line "interrupts" the "RDFGs" can go "TansIdle" or "Partitioned"

     o    When the line comes back, generally the "TransIdle" will resume "Consistent" within a few minutes.

     o    When the line comes back, the 'Partitioned" can also come back to "Consistent" automatically, but may not.

     o    When it doesn't come back it goes "Suspended" (I think).  Then we have to manually resume it:

We do create what we call a Gold BCV (GBCV) on the D/R array:  We split and synch the R2s to the GBCVs every 4 or 6 hours, depending on the RDFG, and we keep those.  We use them for D/R tests, so that we don't have to stop the R1-R2 synch when we're testing D/R.