2010-10-04 01:57 AM - edited 2015-12-18 01:32 AM
i recently went to customer who had 20 storage groups. we sold them window bundle pack and the now pain begins. we installed snapdrive 6.3 and now it takes lot of time to enumrate disks since the guys has 20 luns and each storage group has lun and we have common log location.
i tried netapp support and they kept on asking to give logs , i tried to consildate four storage groups per lun and now snapmanger for exchange doesnt work. i have read best practices and now i am all confused whether to use the volume mount option or not . cant we override the snapmanger for exchange config wizard. ther should be way to do this. the enviroment is exchnage 2007 with windows 2003 cluster
2010-10-04 03:06 AM
please understand that snapdrive has problem when enumeration of disks when they are above 10 and in this case the snapmanager exchange needs lun per storage group and it would actuall be more than 20 . my connectivity is fiber channel.
my patch level should be problem when i reduce the number of luns it works fine.
2010-10-04 06:02 AM
I apologize for the bad experience and will try to help you out here.
1. Its a good thing you opened the support case as this gives us a window straight into escalations if needed.
2. We will need logs to understand where the problems are coming from. It may be a enumeration issue but could stem from something else which the logs will help to explain.
Please send more details on the environment.
Windows 2003 Cluster
Host Util kit?
P.S. I also sent you a direct email.
2010-10-04 08:57 AM
SME iterates on the number of flexvols for transaction log LUNs, so minimizing the number of transaction log volumes and tlog LUNs is a good idea to reduce backup times.
I would be curious on the transport protocol setting being used for SnapDrive. Is the customer using RPC or HTTP/HTTPS?
Setting preferred storage system IP address helps quite a bit. I have found that ensuring that DNS resolution needs to be healthy to help reduce enumeration delays.
The number of nodes in the Windows cluster also affects the enumeration times as well.
Need more information about the environment, than what is being provided to help diagnose the problem.
Can a case number be provided?