I'm trying to implement FLM in our environment and I've a question related to migrating the data of a 4 node cluster which is a VM. Can i implement FLM on a VMware cluster or it has to be a stand alone host used in FLM?
As of 5876.229 we support FLM and VMware for the migration of a RDM or VMFS (though RDM is by far the most useful use-case). It can be migrated from a 5875 or 5773 frame to a VMAX running 5876.229. This can be done using Solutions Enabler 7.6 or Unisphere 1.6.1.
I am rather sure it is supported with clusters, but I am checking on that with engineering now.
Could you please explain your scenario in a bit more detail? You have a 4 node cluster running as VMs on top of ESX host? or you want to migrate data of 4 node ESX cluster?
VMware do not support FLM. Check this article, it only talks about 5875 code, I am not sure if this limitation has been lifted in 5876 code or not?
I'm trying to migrate data out of 6 LUNS to new VMAX and these 6 LUNS are part of Oracle ASM. do we create a new ASM diskgroup for the newer luns or add the LUNS to the existing diskgroup before creating a session?
I don't really know about much about Oracle so I can't say for sure. But FLM will cause the LUNs to look exactly the same on the new array as they did on the old array so the host doesnt really know there is a change. The only change is the logical paths, so if the host is sensitive to that it would be a problem and require changes, but if it was the host wouldnt be supported by FLM. So my guess would be no--you wouldnt have to make a change until the point to unspoof the devices. I am just guessing, you need someone that knows more about Oracle than just being able to spell it
codyhosterman
286 Posts
808
0
Posted September 17th, 2013 07:00
As of 5876.229 we support FLM and VMware for the migration of a RDM or VMFS (though RDM is by far the most useful use-case). It can be migrated from a 5875 or 5773 frame to a VMAX running 5876.229. This can be done using Solutions Enabler 7.6 or Unisphere 1.6.1.
I am rather sure it is supported with clusters, but I am checking on that with engineering now.