Subscribe

SnapManager For SQL Clone Wizard - Slow Enumeration

Hi

I am experiencing issues using NetApp's Snap Manager for SQL Clone Wizard facility.  When I reach the point of selecting a database to clone (or delete) at the point of 'Backup Selection' enumeration of the databases available to clone consistently takes upwards of 30 min's....

My SQL server contains two SQL instances, with multiple databases, stored on multiple mount points - and I am aware that this enironment can contribute to a delay - but delay's of 30 minutes to sinple enumerate what databases are available to clone make this formerly handy tool unusable.

Discussions with my NetApp vendor prompted them to run 'snap list' commands against individual mount points to illustrate how long each individual query takes and thus justify the wait time within the wizard and subsequently suggest how to complete a clone manually using System Manager and Snap Manager but the wizard is a far slicker method if it works...

Can anyone suggest a way toe reduce this delay?  Or does anybody know of any plans afoot to change how NetApp retrieves this list?  I can't believe that this process can't be streamlined somehow.

Any responses welcome!

Thanks

Adam

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

You want to make sure SnapDrive for Windows is working appropriately, as SMSQL depends on SDW. You may want to make sure the time between storage and the host are in sync, check property of disk under SDW to make sure the preferred storage system IP address is set correctly. You may also try to use the latest SDW.

Thanks,

-Qing

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

Hi Qing

Many thanks for your response.

I double checked times on the storage and host and they were 5 minutes apart - I addressed this but it didn't make a difference unfortunately.

When it came to adding a 'preferred storage system IP' there was no setting in there.  I wasn't sure whether to add either the Managment IP or the Public IP but tried both individually and neither entry seemed to improve enumeration speed.

With regards to version level I found it difficult to ascertain the latest version on the NetApp site, but the articles I did find led me to believe that our version of 6.3.4601 was relatively up to date....  Is there a URL detailing version information that you know of that I was unable to find...?

Regards

Adam

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

We are running into the exact same issue!  If no resolution guess we'll need to call support and escalate this afternoon.  Although ours does vary from time to time but not by much.... 15min at the fastest, sometimes 25-30min.

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

Hi ,

What is the version of SMSQL is being used ? Is the same time observed when you smsql clone cmdlet ?

Regards,

Abhishek

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

Same issue is happening in our environment.  We have a database with over 14,000 files and snapmanager times out while enumerating databases. 

Any help with this issue would be appreciated.

Thanks

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

What protocol are you using to connect SnapDrive to the storage system - RPC, HTTP or HTTPS?  If, HTTP or HTTPS, are you using a local account on the controller or a domain account?

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

Was there any resolution to this? Have a customer with the exact same issue of enumeration.

Cheers

Re: SnapManager For SQL Clone Wizard - Slow Enumeration

So, never heard back on the specifics of the settings in this case, but in general these steps will dramatically improve enumeration:

  1. Update the machine to a current combination of Data ONTAP, the Host Utilities Kit (HUK), SnapDrive for Windows (SDW), and SMSQL (see http://support.netapp.com/matrix - check the SnapManager for Microsoft SQL Server area for SQL, SMSQL and SDW versions, and then check the Storage Area Network (SAN) area for ONTAP, SDW, MPIO DSM (MS recommended), and HUK versions)
  2. Use HTTPS authentication for the Transport Protocol and a local user account on the cluster/storage system (do not use a domain account)

Hope this helps,

Seth