What's your requirement? If you are dedicating disks to an application and need optimum performance, MetaLUNs are the way to go.
If you want to take advantage of Thin Provisioning, LUN Compression, FAST VP Tiering, or just want a big pool of disks for multiple applications to share, then Pools can be quite useful.
You may read the following whitepapers for more information:
latest Flare OE is 04.30.000.5.526, which is still available for download on support.emc.com.
Design is a broad question, but I'l give quick some advice about how I would go about it.
The Storage Pool flexibility is more limited on CX4, so you'll need to consider some recommendations such as not mixing drives sizes in a single pool, preferred drive counts and multipliers.
With that in mind I'd do the following when looking at a design;
- Use striped metaluns based on R5 Raid Groups for the most demanding workloads
- Use R1/0 Raid Group(s) for Database Log Luns if warranted (size and perf)
- Use R5 Pools for General workloads.
Consider separate Pools/RG's for isolation and performance protection.
If you have the capacity, consider leave some drives unbound (4+1), so if you're faced with a performance issue, you can 'inject' another group into the pool/rg for a boost.
You haven't mentioned Fast VP, and without multiple drive types, it will be of no benefit as I'm fairly sure there is no intra-pool tiering in CX.
The best practice documents linked above lists all recommendations and restrictions.
This is general advice though, without knowing your workload requirements.
dynamox
9 Legend
•
20.4K Posts
0
December 18th, 2014 11:00
you define hot spares manually on CX4/VNX1 arrays.
Roger_Wu
4 Operator
•
4K Posts
1
December 14th, 2014 18:00
What's your requirement? If you are dedicating disks to an application and need optimum performance, MetaLUNs are the way to go.
If you want to take advantage of Thin Provisioning, LUN Compression, FAST VP Tiering, or just want a big pool of disks for multiple applications to share, then Pools can be quite useful.
You may read the following whitepapers for more information:
http://www.emc.com/collateral/hardware/white-papers/h5773-clariion-best-practices-performance-availability-wp.pdf
http://www.emc.com/collateral/software/solution-overview/h2197-vmware-esx-clariion-stor-syst-ldv.pdf
https://ecn-dev.emc.com/servlet/JiveServlet/download/481038-18774/H1416-emc-clariion-intgtn-vmware-wp.pdf
http://www.emc.com/collateral/software/white-papers/h6131-implementing-vp-with-clariion-celerra-vmware-infrastructure-wp…
brettesinclair
2 Intern
•
715 Posts
1
December 14th, 2014 20:00
latest Flare OE is 04.30.000.5.526, which is still available for download on support.emc.com.
Design is a broad question, but I'l give quick some advice about how I would go about it.
The Storage Pool flexibility is more limited on CX4, so you'll need to consider some recommendations such as not mixing drives sizes in a single pool, preferred drive counts and multipliers.
With that in mind I'd do the following when looking at a design;
- Use striped metaluns based on R5 Raid Groups for the most demanding workloads
- Use R1/0 Raid Group(s) for Database Log Luns if warranted (size and perf)
- Use R5 Pools for General workloads.
Consider separate Pools/RG's for isolation and performance protection.
If you have the capacity, consider leave some drives unbound (4+1), so if you're faced with a performance issue, you can 'inject' another group into the pool/rg for a boost.
You haven't mentioned Fast VP, and without multiple drive types, it will be of no benefit as I'm fairly sure there is no intra-pool tiering in CX.
The best practice documents linked above lists all recommendations and restrictions.
This is general advice though, without knowing your workload requirements.
rmcnew-orangelake.com
27 Posts
0
December 18th, 2014 06:00
If I create storage pools, should I create hot spares first or does the storage pool creation process do this?