ONTAP Discussions

Way of utilising existing Snapmirror job for longer retentions?




I am really hoping someone can advise me on this as after week of setting up SV its not working the way I am expecting. 


The issue I am having is I have created a SV relationship on my destination filer but the source copy is a snapmirror destination volume. I then want to create more retentions from that Snapmirror destination volume e.g. 1 month as the SM job only retains 2 days of backups between the source and destination filer. All of our flexvols are luns presensted to a VMware farm so the data is VMFS files only. 


The reason why I am doing this is I want to utilise the already present snapmirror relationship and so dont want to create another snapvault relationship back to the primary because it means another baseline backup would need to be performed which would take a very long time. As the snapmirror has already done this for me, I want to just create longer retentions from the snapmirror backups but its not working as expected?


I created a new flexvol on the destination filer where all the SV jobs would go to and it has been saving snapshots however any new data change updates from the primary filer to seconday are not getting copied over, I have proved by creating flexclones of the volumes at both source and desintation, on the source I can see the file changes made when I bring up another instance of the live VM and inside the VM I can see the changes, if I do the same on the desintation filer I dont see those changes.


Snapvault is performing a backup on the desintation side from the snapmirror destination volume.

My questions are is this possible? If so how would I set this up? I am not performing any preserve snapshots btw on both filers, also I was under the impression that I would need qtrees only on the destination volume? as my source volumes dont have qtrees is that going to be ok? will the cause issues withe SV? 





According to documentation in mirror-vault cascade only snapshots with label sm_created are replicated to vault secondary.


I am using Data ONTAP 7-mode. 


Am I right in saying that what you mention only works with Clustered ONTAP?





In 7-mode you configure primary SnapVault schedule on SnapMirror source (assuming VSM). SnapVault secondary does not care, where snapshots come from - it simply fetches snapshots with specific name(s). Of course you need adjust schedules so SnapVault transfer runs after SnapMirror transfer in complete.


Thanks for your reply.


I am aware that I can create from snapvault jobs from Destination to source and I have done this for some new flexvols I have created which didnt utilise a snapmirror relationship.


However I do have existing Snapmirror relationships already in place which are working between our source and destination filer. If I create new snapvault jobs and target it back to the primary filer, wont it need to do another baseline transfer first as the destination volume will be dedicated for the snapvaults repository? If this is the case this is what I am trying to avoid as the baselines transfers will take a very long time.


Thats why I am hoping for the existing snapmirror volumes I can just perform snapvaults of those for volumes where I need a longer retention.


Is there a way I can achieve this without rerunning new basline transfers?


hope that make sense?

I am afraid I do not quite understand what you try to do. Could you outline your config, like "you have volume A that is snap mirrored to volume B and you want to SnapVault B to C" or similar. At least, that is how I understood you initially.




Ok maybe I haven't explained it properly.


Here is an entry for one volume that is snapmirrored across from one filer to another.


FASPROD12:NETAPP_prf2_esx_w2k12_sas_01 FASDR12:NETAPP_prf2_esx_w2k12_sas_01 kbs=80000 * 7,13,19 * 0,1,2,3,4,5,6


FASPROD12 - is our source and Production filer. At primary site.  

FASDR12 is our destination and DR Filer at DR site


Now what I want to do is create a snapvault relationship on the DR/destination filer of the volume NETAPP_prf2_esx_w2k12_sas_01 so that I can create longer retentions as the snapmirror only has a retention of 2 days.


Below is the snapvault job I have created on the destination filer, it has performed the initial baseline transfer ok but I am having trouble it creating further snapshots from this, the last snapshot it created was the 27th of last month and I dont fully understand why it has suddenly stopped.




Hope this clarifies things a little







Now what I want to do is create a snapvault relationship on the DR/destination filer of the volume NETAPP_prf2_esx_w2k12_sas_01

I'm sorry, I do not understand it. SnapVault is created between source and destination. What is source filer and qtree and what is destination filer and qtree?


Hi again


Ok on the source side we dont use qtrees at all. The snapmirrors are all volume snapmirrors and where we are using Snapvault they are all "non-qtree snapvault relationships". On the destination side though there needs to be a qtree so there is a qtree for the snapvault volume destination so in the case of the example I gave earlier it is set out like this:


Source - CENWBFASDR02:/vol/NETAPP_prf2_esx_w2k12_sas_01/-

Destination - CENWBFASDR02:/vol/SVBackups_NETAPP_prf2_esx_w2k12_retain01m_sas_01/q_NETAPP_prf2_esx_w2k12_retain01m_sas_01 (q signifies its a qtree)






Sigh ... you again give different names as before, so I only can assume that CENWBFASDR02 means the same as FASDR12 in your previous post. In this case it is as I already told you - you need to setup SnapVault snapshot schedule on FASPROD12 (taking your example). These snapshots will be replicated to FASDR12 by SnapMirror and then picked up by SnapVault.


Yes sorry FASDR and CENWB fas are the same thing.


So are you saying if I create the snapvault from the destination and point it to the source when snapvault runs it won't do another baseline copy, as that was my concern.


Will snapvault know there is already a snapmirror relationship and only copy the snapshot data and not the whole baseline? From reading your last comment thats how I am understanding it will work... Is that correct?


Thank you for perservering.


So are you saying if I create the snapvault from the destination and point it to the source when snapvault runs it won't do another baseline copy, as that was my concern.

There seems to be some misunderstanding. Assuming destination and source mean SnapMirror destination and source, you do not "point SnapVault to SnapMirror source". You create SnapVault relationship with source == SnapMirror destination and leave it running. Because SnapMirror destination is read only, you cannot create SnapVault schedule there; so you create SnapVault schedule on SnapMirror source and it gets replicated to SnapMirror destination.


All of this is described in NetApp documentation.




Ok let me give that a try and see how I get on.


I have been creating all my SV jobs from the destination filer and not the source filer.





When you say "All of this is described in NetApp documentation."


Do you mean this document?


Data ONTAP® 8.2

Data Protection Online Backup and Recovery Guide

For 7-Mode



Yes, this is the manual I mean.




Ok I have created a new SV relationship from the source filer


Having read the documentation a fair few times over the last month I think you right about the misunderstanding part!


I still dont understand how this is going to work and how the destination filer will get to know about the extra snapshot copies based on the retention policy that I have created on the source. Also if I have created this relationship on the source won't these snapshots retain on the source filer as this is not how I want it to work as I want our DR filer to keep the retentions and not staore these on the production filer as this will consume unncessary space. the maximum amount of snapshots I wanted to retain on the production filer is 7 days but on the Dr filer I want to keep up to 1 month.


When I created the relationship on the source it automatically made the transfer type to "transfer", I was under the understanding that transfer was meant to be used from the destination filer so it can get updates from the source?

Do I need to create a schedule on the destination filer with the same schedule plan for the two to marry together?


Thank you