2010-08-06 12:00 AM
We try to set up FAS2040A for VMWare environment with FCP protocol. One FC port on each controller is used for disk shelfs. We have also 2 brocade FC switches.
We have only one FC port on controller left. We can only connect one filer to the first switch and other filer to the second switch. With that topology we can not achieve full redundancy if one switch fails.
Is possible to make multifabric active/active configuration with this topology? This topology was supported on FAS270c platforms (dual fabric). (http://now.netapp.com/NOW/knowledge/docs/san/fcp_iscsi_config/config_guide_73/GUID-A17C624B-BE9E-4CC0-B3A1-DC70ECE3992B-low.gif)
2010-08-06 12:49 AM
No, that’s incorrect. If one switch fails, LUN access is retained via partner connection. This is the same as was on FAS270C. You have to set cfmode to single_image, but this is effectively the only cfmode supported today.
And, BTW, FAS270C dual_fabric was never active/active ☺
2010-08-06 01:13 AM
So we have two switches and two partners (2040a). We have only one available FC connection on each FAS, because we already connected disk shelves.
We connect first FAS partner to first switch and second FAS partner to second switch and use single_image cf_mode.
so if i understand you right...If one switch fails, LUN connectivity is retained internaly via partners connection?