When will Protection Manager support volume-based SnapVault?

[dfaul1234]

I have a customer with many TBs of VMware datastores within flexvols without qtrees.  When will Protection Manager support volume-based SnapVault?  We are implementing SnapVault for long-term VM retention and could really use Protection Manager for backup management.

Any plans to tightly integrate SMVI, SnapVault, Protection Manager, System Manager etc.?  The options currently available to protect different data can be very confusing to customers.  I envision a unified/combined management solution that will integrate all the current "snap" technologies so customers have a single view of their data and protection policies.

Thanks,

Dave

[adaikkap]

Hi Dave,

     Even now Protection Manger supports Snapvault of entire volume(irrespective of qtree is there or not.)

What protection manger does is that, it treats those data that is  inside the volume but not the qtree as non-qtree data represented by

<filername>:/<volumename>/<qtreename> in case of non qtree the qtree name is replaced with "-"

Example.

filername:/vin1_root/-

So you can still protect your ,TBs of VMware datastores within flexvols without qtrees using protection manager.

Regards

adai

[dfaul1234]

Thanks for the quick reply.  The other question was will there be better integration with Protection Manager, SMVI, SnapVault etc. in the up-coming release of Protection Manager?  How can we help customers integrate Protection Manager, SnapVault and SMVI now or will this be better integrated in the near future?

Thanks,

Dave

[smoot]

Yes, the next release will largely merge the two products.

[dfaul1234]

Does Protection Manager support importing already created volume-based snapvault relationships?  It doesnt seems to list them under External Relationships.

Thanks,

Dave

[smoot]

Sadly, we have no plans to support whole-volume SnapVault. As you have discovered, we don't even discover the relationships.

[cui]

Hi Dave,

Good questions.

Integration of SMVI, SnapVault, Protection Manager is planned for mid 2010.  If you need additional information, please contact me offline.

Thanks,

Cindy

[cui]

No.  Protection Manager does not recognize WVSV.

[dfaul1234]

OK, it doesn't reconize them but it does let you create vol-based SV relationships from scratch and manage them via PM right?

Thanks,

Dave

[smoot]

No, our monitor doesn't recognize the relationship on the storage system, so it doesn't show up in our database.  Since it's not in our database, ProtMgr won't allow you to import it.

[adaikkap]

yes. it doesn't allow you to import existing relationship but does allow creating from scratch and managing the same.

[cui]

There is a misunderstanding on what does volume based SV support mean.

1. You can put a whole volume into protection manager and PM will create the correct SV relationship.  It does not use the ONTAP WVSV mechanism.  This is what you need for the usecase you described in the original post.

2. Some customers use WVSV, which is ONTAP technology to get around the SV performance issues when there are thousands of qtrees in a volume.  This is not the usecase you have. Protection Manager does not support this configuration today.

Thanks,

Cindy

[dfaul1234]

Please explain the acronym WVSV.  I think there is a confusion here....

[cui · Accepted solution]

Whole Volume SnapVault.

[smoot]

Let me make sure I'm clear.  If you already know all this, then excuse me for wasting bits.

There are two ways to do what you're trying to do.

One way is to set up a "normal" SnapVault relationship where the source is "-", sometimes called "qtree0".  This represents all the data not in qtrees in your volume.  Since you said you don't use qtrees, this essentially means all the data contained by the volume.  This approach is 100% supported by Protection Manager.

The second way is "whole-volume SnapVault".  In this case, the source is the volume itself, not qtree0, so you'll see a source path like "filer:/vol/mvol", not "filer:/vol/myvol/-".  No DFM-based product understands whole-volume SnapVault.

Protection Manager tries to finesse this distinction.  You can add a volume to a dataset and we'll ensure all the data in the volume is protected.  Today, we do that by building SnapVault relationships from all source qtrees and one additional relationship for qtree0.

If you already have whole-volume SV relationships, I'm not sure if there's a way to convert them into qtree0-based relationships.  If not, then you won't be able to get Protection Manager to work with them.

[dfaul1234]

Thanks for the great feedback Cindy and Peter.  This was a great help.

Dave

[kjdonnelly]

Hi,

I am wondering if there is an update to the support for DFM 4.0 to support the WVSV as mentioned in this forum.

The situation I have is that a customer has created cifs volumes with a large amount of qtrees delimiting their smb share point from these qtrees within the volume, for example we have:

/vol/myvolume/busfinqt

                    /busfinblah

                    /busfinblahblah    

                    /blah

                    /blahblahetc..

(the part from the last / are qtrees)

There are approximately 250 qtrees in multiple volumes that they are looking at replicating to a snapvault location, prior to this the individual qtree method was tried and tested although the performance impact (even with Nearstore enabled) caused them to drop that design and go back to the drawing board.

The ideal end point would be for DFM/NMC (Netapp Management Console) to recognize the wholevolume snapvaults so I can import it into a mirror and backup policy, I am talking about the policy which has a different destination for each snapmirror / snapvault within the NMC rather than a cascade source => snapmirror=> snapvault relationship. It basically looks like this:

                   /  SnapMirror

Sourcevol   / 

                 \

                   \  Snapvault

Otherwise, the options are to live with the external management of snapvault at the WVSV level, or redesign the volumes with a higher level share and ACL the subdirectories off below i.e.:

/vol/myvolume/topshareqt/sub1

                                     /sub2

                                     /sub3

/vol/myvolume.topshare2qt/sub1

                                       /blah..

Any info or guidance appreciated!

Thanks

KD

[adaikkap]

Hi Kd,

     You can add your /vol/myvolume to the dataset having Backup and mirror Policy, as primary member.

PM will take care and create 250 Snapvault relationships in the Backup filer and create one Snapmirror relationships in the mirror node.

Any further volumes added in the /vol/myvolume is automatically detedted by PM and SV relationships are created.

But still PM does not supported the whole volume snap vault of /vol/myvolume backedup  to a single qtree in the destination volume thus creating one SV relationships.

Hope I clarified your doubts.

Regards

adai

[emollonb2s]

I think what you mean, but then i will have kindo of DR policy, i just want to retain 10 days of my virtual machines on local and then retain a year on my secondary node (backup), there is any way to do it using Snapvault?

If I'm using snapvault integrated with SMVI i could have a consistent snapshots of my virtual machines, if i don't use smvi the all the snapshots from PM will not be a consistent snapshots because i have to call smvi to take the snapshots before the replica starts.

Any solutions?

Thanks!

[adaikkap]

I think what you mean, but then i will have kindo of DR policy, i just want to retain 10 days of my virtual machines on local and then retain a year on my secondary node (backup), there is any way to do it using Snapvault? 

Have different retention in the primary node and the secondary node.

[root@lnx~]# dfpm policy node get 61
Node Id:                    1
Node Name:                  Primary data
Hourly Retention Count:     2
Hourly Retention Duration:  86400
Daily Retention Count:      2
Daily Retention Duration:   604800
Weekly Retention Count:     1
Weekly Retention Duration:  1209600
Monthly Retention Count:    0
Monthly Retention Duration: 0
Backup Script Path:
Backup Script Run As:
Failover Script Path:
Failover Script Run As:
Snapshot Schedule Id:       46
Snapshot Schedule Name:     Sunday at midnight with daily and hourly
Warning Lag Enabled:        Yes
Warning Lag Threshold:      129600
Error Lag Enabled:          Yes
Error Lag Threshold:        172800

Node Id:                    2
Node Name:                  Backup
Hourly Retention Count:     0
Hourly Retention Duration:  0
Daily Retention Count:      2
Daily Retention Duration:   1209600
Weekly Retention Count:     2
Weekly Retention Duration:  4838400
Monthly Retention Count:    1
Monthly Retention Duration: 8467200

[root@ln~]#

If I'm using snapvault integrated with SMVI i could have a consistent snapshots of my virtual machines, if i don't use smvi the all the snapshots from PM will not be a consistent snapshots because i have to call smvi to take the snapshots before the replica starts.

You can use NetApp Managebility SDK, to register the snapshots taken by SMVI as a backup version in PM, and also create an Application Dataset in PM so that PM will only transfer the registered consistent snapshot to the secondary instead of taking its own

snapshot.

Regards

adai

[emollonb2s]

This means that i have to develop my own aplications to manage my snapshots from smvi to transfer it to PM?, i think it's too much for me man! can you provide me any useful information to do that?

Thanks very much Adaikkappan for your help, but for now i think i'm going to use SV-SMVI for a while.