Subscribe

SDW/SME slow to perform out of band operations when large number of luns connected to windows host

I'm encountering an issue with a new Windows 2008/Exchange 2007 SCC cluster where Snapdrive 6.02 takes a long time to perform out of band operations such as disk enumeration, snapshot list,snapshot create.   I believe this is causing SME backups to take a long time and will affect our RPO.  In this environment we have a four node active/active/passive/passive cluster with two Exchange virtual instances. Each Exchange virtual instance has 25 storage groups so we have 50 luns (separate luns for db/log for each storage group) x2 for a total of 100 luns connected to this SCC cluster. I've opened a case with support #2000680173 and I'm sending in the OntapWinDC output from each cluster node.  I'm using https as the SDW transport protocol, there are no production mailboxes on this cluster yet. I ran an unverified SME backup on one storage group and while the job was successfull it took about 20 minutes to complete.

The environment consists of 4 Windows 2008 hosts with Data Ontap DSM for Windows 3.2.R1, SDW 6.02, SME5.0, host utilities 5.1 connected via FCP to two FAS3170A clusters.

The filer are serial #s 70000081, 70000082, 70000085, and 70000086.

Has anyone else seen this behavior?

Thanks,


Ryan Rothert

Re: SDW/SME slow to perform out of band operations when large number of luns connected to windows host

Part of the SDW 6.0.2 Release notes.

HTTP and HTTPS slow when using domain account

Re: SDW/SME slow to perform out of band operations when large number of luns connected to windows host

Ah....very good catch....curious to know if any difference if you change to RPC authentication?