oops, i thought we were talking about initiator cascaded groups. It does not have to be cluster allocation. For example i have an Oracle system where i have data file systems and archive log file systems. I would like to use FAST for my data file systems so then can go up and down between my T1 and T2 tiers but keep my log file systems on the same teir (because of its sequential workload not very beneficial to be on EFD tier). So before code 76 i would have to create two different storage roups, two different masking views, one would be for my oracle data TDEVs and another would be for my log TDEVs. Now i can have just one masking view, one cascaded storage group.
Thanks Saurabh , I can see your point about the usage of cascaded SGs in FAST Policy distribution.Its remarkable.. But i am still in dark about the role of same in a Cluster allocation. Could you please elaborate bit more?
Hi Dynamox , i think you are talking about the Cascaded IG.Yes .. you are right about it.But i was just wondering about the purpose of Cascaded Storage Groups(SG) , which i saw recently in SE 7.4 guide.
when you have any type of cluster where you need to share devices, yet also have the ability to present non-shared devices to cluster members. You would create cascaded initiator group that would be member of your "shared" LUNs masking view and then use individual initiator groups that will be part of your "individual LUNs masking view.
Rajesh, one example is what dynamox has described where you would need it for clustered host. Other significant use of Cascaded SGs will be in FAST or FAST VP. This will give you flexibility to create one child SG and associate with a fast policy and other child SG to another fast policy and parent SG will be in masking view.
Can I remove devices out of one child group and put them in another non-disruptively? The devices are still in the parent group and unchanged. Will this impact the masking to the host?
dynamox
11 Legend
•
20419 Posts
•
87439 Points
2759
1
Posted July 23rd, 2012 06:00
oops, i thought we were talking about initiator cascaded groups. It does not have to be cluster allocation. For example i have an Oracle system where i have data file systems and archive log file systems. I would like to use FAST for my data file systems so then can go up and down between my T1 and T2 tiers but keep my log file systems on the same teir (because of its sequential workload not very beneficial to be on EFD tier). So before code 76 i would have to create two different storage roups, two different masking views, one would be for my oracle data TDEVs and another would be for my log TDEVs. Now i can have just one masking view, one cascaded storage group.