storage configuration per host and vm

[dbdxjq]

   hi,

i got a few questions about the storage.

1, i can see total provisioned LUN space is less than the vm disks space, and we are using thick disk type. for example, the size of created LUN is 420GB, and provisioned space is 380GB, while i can see the size of disks located on this storage is 620GB. we are using both thick lazy zeroed disk and thick eager zeroed disk, but surely no thin disks.

2, is there a bottleneck for storage? could the storage usage hit 100%? if it reach 100%, is there any potential risk? BTW, i assume we dont need snapshot.

3, how much  is the largest LUN size that vsphere 5.1 support? and is there a recommended value for the size of LUN？

4, how many vms can be located in one LUN in vsphere 5.1? and is there a recommended value for this if i have enough resources?

thanks!

[vmroyale]

1 - where are you seeing these sizes reported?

2 - storage can definitely hit 100% and it probably won't be a good thing if it happens. Check out kb1003412 for more info.

3 - you can have a 64TB VMFS volume or RDM (physical mode) in 5.1 - RDM (virtual mode) is limited to 2TB minus 512 bytes. The recommended size depends and will vary.

4 - This too depends on a variety of factors. Check out VAAI and the Unlimited VMs per Datastore Urban Myth for some good info on this.

What more can you tell us about your storage, pathing and workloads?

[dbdxjq]

thanks for your kindly answer.

1, i can basically have the size on datastore view via vsphere client and count the size of vm disks, and then compared the result. does this matter with thick lazy zeroed disk?

2, this is already enough for me. thanks again!

3, i understand this depends. but could you tell me more about the conditions which will impact the value? or how am i able to identify the size for each LUN to get the best performance?

4, if we goes with one vm per LUN, can i say we will have the best performance of the vm?

we have 11 hosts, one hundred or more vms in hand. and 11 hosts are seperated into 2 clusters. and its hardly to define the workload for vms, cause different applications are running on different vms. but all of shared storage are with fibre channel. we have more than 60 LUNs, big LUN with 3TB is located more than 15 vms, small LUN with  50GB is located probrably one vm. and totally we have created more than 50 LUNs

what i'm trying to do is to figure out the proper size and proper vms for each LUN.