2014-02-12 11:58 AM
First post here, hope its in the correct place!
We have recently done a failover test of one of our servers and i have a question off the back of it. I will detail our procedure below so you have an understanding of what we are doing. All commands are run in powershell with the dataONTAP module imported.
on source server
we use sdcli to create a snapshot and update the mirror.
Check that the mirror has finished transferring
On dest server
use sdcli to connect disk to initiators of the server
So my question, once the disk is in windows and the product team have tested it and want to fail back we have to run lun clone split status and wait for that to complete, before mirroring back. I wold expect to have to do this if i was splitting a flexclone from the parent volume but why does this happen when connecting to the LUN in a broken off snapmirror destination???? what is being done here and what triggers it? is there a way to mirror until it has finished? also how big will the top snapshot grow as the source volume is 4.5tb.
just as a fyi the @snapmir@..... snapshot says busy,lun and the one above it is growing in size.
sorry about all the questions, this is an area of netapp that i dont understand.
2014-02-13 01:10 PM
Do you need the lun split or can you destroy the lun after the test? Destroy instead of split and the busy will clear...or use the sis clone feature to use the dedup engine for the clone to not lock a snap.