Guys, your advice would be appreciated on this topic...
I've recently installed a 6 server ESX Host Cluster & connected these to a new shared storage platform over iSCSI.
I'm now in the process of starting to do P2V conversions.
All physical servers have data stored locally, as there has been no shared storage solution implmented until now.
Has anyone implemented tiered levels of storage presented to a VMWARE Cluster:
I'm considering implementing 3 tiers as 3 different VMFS volumes within VMWARE as shown below:
High Cost / High Performance / Low Capacilty - 15k 300Gb SAS
Med Cost / Med Performance / Med Capacity - 10k 400Gb SAS
Low Cost / Low Performance / HIgh Capacity - 7.2 1TB SATA
What way would you configure a VM within these VMFS volumes?
E.g. For a vm containing a high end Database with high levels of I/O transactions i would consider putting the O.S hd on the middle tier & the data hd on the high tier.
E.g. For a vm containing an archive server, the O.S would be placed on the middle tier, with the data placed on the low tier.
Has anyone implemented such a solution?
If so have you notices any improvements in performance etc?
Also has anyone implemented ILM across such a solution, with a single VM having Virtual Disks on all 3 tiers, automatically moving data between tiers once it reaches predetermined tresholds?
If so what software have you used?
What software was used to do the pre storage analysis of the data stored locally on physical servers, in order to determine how much storage was required.
I understand this is alot to ask, but it something that I am actively looking at & any help/suggestions would be appreciated,
Regards,
