Guide for instant cloning of mssql databases

Frank,

I would start by making sure that your array you are using the clone is on the list of supported Arrays.

https://documentation.commvault.com/11.24/expert/33073_supported_storage_arrays.html

Creating an Instant Clone for SQL Server

https://documentation.commvault.com/11.24/essential/87384_creating_instant_clone_for_sql_server.html

When you say slow, what are you comparing that to? Usually a array (command line) clone depending on the volume size can take some time too.

Hi dude,

thanks for your fast reply !

We use both Tintri and Hitachi Vantara arrays and also implemented Intellisnap on those arrays for CommVault VM backups and that works like a charm with minimal impact to our vSphere hosts

We also found the doc on cloning the database but before you can do that you have to create an application aware sqlserver database backup; we did that on both array types with Intellisnap and basically it works but it takes quit some time before we're able to clone the database backup-ed in the first step and also the clone takes considerable time. So compared to the classic method (using standard backup and restore) we don't seem to win much time.

We presumed we only needed the Intellisnap copy without running a backup copy step like we do in our VM backups with Intellisnap because that removes the snap that we need for the clone

We also want to create the standard database backups beside this cloning stream. So we're looking for more detailed information.

We could create a support case of course but first wanted to find out if there's any experience in this field within the community : )

You bring a great point and interesting discussion. I have tried cloning for Oracle which “works” but I also encountered some challanges along the way. Way too many settings, not to mention, the communication between Commvault and where the volume resides, mounting the cloned database, space soooo….

Ultimately we decided not to use Commvault for clones and instead just run a regular array clone whenever needed (not often).

My guess is that someone from CV will see this and provide some additional info. Maybe @Mike Struening  

Thanks for your update;

We cloned DB2 mainframe db's on EMC Symmetrix with the array’s TimeFinder options twice every day flawlessly for years a long time ago . . .  : )

And Tintri also has mssql cloning options, but we just tried to put CommVault in the orchestrator role since all functionality is there: Intellisnap, app aware agents, mssql integration and array transparant options to make this work !

So it would be great to have some help on this from CV support . . . 

Hi @frankvaniersel .  hope all is well!  Sorry for the delayed reply, I was on vacation last week.

I can tag in some people here to comment, though I’m leaning towards opening a support case for the performance concern (and sharing the incident number here).

Let me see who I can get to chime in here (I know a few people off the top of my head).

Hello @frankvaniersel - I would agree with Michael. While often the speed is related to different factors (data size, infrastructure) if you wanted to investigate the performance, a Support case might be best suited so we can get details and advise. 

Hi Mike and Ron,

i will go ahead and open a support case and share the case number in this call so we can share all information for other users on this topic : )

cheers,

frank

Case is created:  #220209 229

we included a test scenario and will include the corresponding job logs in that case

Hi,
thanks for the update

so this will not work in our setting; and good to know that : )
we'll investigate other options

We have similar use-cases but until now the current implementation just satisfy the needs of our customers. If I recall correctly Commvault will make the DB available on-the-fly and will start using local disk space as a cache for read/write once you start using the clone, but this impacts performance and the use-cases for using the database cloning functionality. As far as I know it is targeting quick tests on databases and or ability to perform quick lookups/verification, but less for full application level testing. Besides use-cases, but in relation to them I miss some options that:

• Allow the user to not having to enter a clone reservation duration. E.g. option to leave the clone intact without end date. 
• Ability to perform a full recovery (yes this could be done using a regular backup/recovery) but while leveraging a local recovery cache that allows future updates to be "quick” by leveraging incremental recovery. 

Especially in more enterprise environments and cloud environment in where environments like production and test (and others) are separated due to security reasons (or others). in this it becomes a challenge to deliver this ability to the business, but as Commvault is already present it would perfectly make sense to have it steer this all. 

To be more clear our use-case entails DBA'ers doing self-service aoo-level recoveries of data out-of-place in a fast and secure wat, taking into account our stack.

Maybe there's a way to have CommVault as orchestrator for these and our use cases deploying NetApp snapcenter integration ? (if you have a NetApp array in place : )