Provisioning behavior issue

Let's take the following scenario to illustrate the problem

1. create a volume using ontap cli --> vol create voltest aggr0 1g

2. generate a new dataset and put the volume "voltest" as primary

3. apply a provisioning policy and make sure the "custom label" is set with the same volume name ("voltest")

4. provision a new share

At this stage Oncommand will generate a new volume called "voltest_1" and put a new qtree in it.

The intended scenario would be Oncommand to resize the existing volume and add a new qtree in it.

We were able to have the proper behavior by generating a DR of the volume (protection policy, backup-dr, then failover dataset, then remove protection policy).

After this step the provisioning will reuse the existing volume.

Re: Provisioning behavior issue


Provisioning Manager doesn't take into consideration of a volume, which is not created through Provisioning Manager, to reuse (like a qtree creation). This is a limitation in Provisioning Manager for now.

Provisioning Manager allows the volumes, which were created through any other interface(Ontap/Filerview/SystemManager), to be a member of dataset to make protection relationships but not for reuse.

I hope this helps you..!

Warm Regards

Sivaprasad K

Re: Provisioning behavior issue

There are no ways, with dfmapi, to integrate the volume in this dataset that I can provision storage in the same object ?

That will be very helpfull.



Re: Provisioning behavior issue

Just for your information, I had already reported this behaviour issue during the Prague Inside in October 2010 directly to the software developers during a presentation.

Please understand that when doing an implementation of provisioning manager to a customer, it is really a problem not beeing able to provision from existing volumes.

Also the Netapp message is "no need to migrate" from Netapp to Netapp when upgrading heads, which is great, however you have to migrate data when implementing provisioning manager if you wish to use only one volume per dataset. This is difficult to justify to customers.

Netapp is seen as a software company, I guess such change request should not be too tricky to implement.

Thanks in advance