UNSOLVED

dvkadam

updated

15 years ago

D

dvkadam

2 Posts

0

533

October 13th, 2011 14:00

RDF split status needs to be failback

we did failover 1 week back
the client wanted to make both sites RW enable after failover
so i did make original R1 as RW enable...as soon as i did it the RDF pair state became SPLIT

now i want to failback but before which i have to do Update and sync all the changes that were done is last 1 week on Original DR site to original R1 site.

we have RDF pair in Split state as of now...
how do i proceed....
should i restore from current state or just do fail back

please advice

  • 263

    0

    Posted October 13th, 2011 16:00

    Just issue --

    symrdf -g DGNAME failback

    R1 side wil  be RW state and R2 will be WD state.Also R1 will start updating from R2.Check with the query command for the no of invalid Tracks...

    symrdf -g DGNAME query -i 5

  • 263

    0

    Posted October 14th, 2011 01:00

    few questions i have:

    1) Why did you make R1 read write?

    2) Have you written anything on the R1 when you were in Failed over state?

    3) If yes, then both sides should have invalid owed to respective sites.

    4) In this case, you wont be able to failback until you clear the local invalid on R1.

    5) Do a symrdf query on the group to confirm the invalids on both sides.

    6) In simple words, you have to choose which site data you want?

    7) If R2 site data is valid, then make R1 write disable (this will change the rdf state to failed over).

    8) then try update r1 command.

    9) if it works then wait untill the tracks come down some good number and then do a failback.

    10) if update r1 fails then engage PSE lab to clear the invalid on r1.

    regards,

    Saurabh

  • dvkadam

    2 Posts

    263

    0

    Posted October 14th, 2011 08:00

    thanks for everybody response

    the R1 was made WD and the RDF pair state changed to Fail over

    once it was fail over i coupld update the R1 with R2 changes with force flag...

    now i am ready for failback which will save good amount of time.

    Thanks Again