Hi Luc,
i am thinking to do initial vsan cluster design calculation using powercli .can you suggest if this has already been done or if we can work on creating something like this.
Don't you see the 'Additional Cluster Information' part?
-----------------------------------------------------------------------------------------------------------------------------------------------
Was it helpful? Let us know by completing this short survey here.
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
Not too sure what you want to calculate.
Do you mean something like the sizing exercise like Cormac mentioned in Virtual SAN – Sizing Considerations
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
thnaks i am checking this .however i am going to share in the form of powercli script for your suggestions.
i thought of doing in following way though its not complete yet but if you can have a look and suggest what all needs to be added to get precise storage memory cpu requirement of esxi .
requirement:
is to migrate vms on fc backed cluster to vsan and right now in below script we are calculating resouces so that we can size vsan cluster .
write-host "vsan capacity Planner test" -ForegroundColor Green
$source_fc_cluster=read-host "source cluster name"
$type_of_destination_vsan=read-host "please provide type of vsan"
$cl=get-cluster $source_fc_cluster
if($type_of_vsan -eq "hybrid")
{
write-host "doing calculation as per inputs"
write-host "since vsan cluster is hybrid raid option available is mirroring"
write-host "input can be provided through csv file also"
$vms=get-vm -Location (Get-Cluster -name $source_fc_cluster)
$fragments_storage_thin = @()
$fragments_memory = @()
$fragments_vcpus = @()
$fragments_storage_thick = @()
$Fragments_swap_space = @()
##
#$harddisks_thick=get-vm $vms|Get-HardDisk | Where {$_.storageformat -eq "Thick"}
foreach ($vm in $vms)
{
$mem_vm=$vm.MemoryGB
$reservation=$vm.ExtensionData.ResourceConfig.MemoryAllocation.Reservation
$swap_space=$mem_vm-$reservation
$vcpus_count_vm=$vm.NumCpu
$fragments_memory += $mem_vm
$fragments_vcpus += $vcpus_count_vm
$Fragments_swap_space += $swap_space
$harddisks_thin=get-vm $vm|Get-HardDisk | Where {$_.storageformat -eq "Thin" }
$harddisks_thick=get-vm $vm|Get-HardDisk | Where {$_.storageformat -eq "Thick" }
foreach ($h in $harddisks_thin)
{
$capacity=$h.capacityGB
$percent_growth=$capacity*.20
$estimated_capacity=$capacity + $percent_growth
#$estimated_capacity
$fragments_storage_thin += $estimated_capacity
}
foreach ($h in $harddisks_thick)
{
$capacity=$h.capacityGB
$fragments_storage_thick += $capacity
}
}
$sum_storage_thin = $fragments_storage_thin -join '+'
$s1=Invoke-Expression $sum_storage_thin
$sum_storage_thick = $fragments_storage_thick -join '+'
$s2=Invoke-Expression $sum_storage_thick
$sum_memory = $fragments_memory -join '+'
Invoke-Expression $sum_memory
$sum_vcpus = $fragments_vcpus -join '+'
Invoke-Expression $sum_vcpus
$total_storage_withoutoverhead=$s1+$s2
$sum_vcpus = $fragments_vcpus -join '+'
Invoke-Expression $sum_vcpus
$sum_swap = $Fragments_swap_space -join '+'
$swap=Invoke-Expression $sum_vcpus
write-host "storage requirment considering raid 1, till now ....in GB" -ForegroundColor Gray
2*$total_storage_withoutoverhead+$swap
I don't think that a sizing exercise is as straight-forward as this script seems to imply.
There is the workload aspect of the VMs to take into consideration as well.
For example, a VM that uses only 10% of the allocated CPU and memory resources, shouldn't be taken at its configuration values.
So, imho, you also would need to look at the workloads (performance statistics) over a period of time.
What is the average consumption? But also which maxima needs the environment be able to support.
And since you want a growth margin, you will need to add spare capacity for the near future.
How fast, and at which rate, does the environment grow (again statistical data).
The official VSAN Sizing calculator also looks at your workloads.
And there are people who have created spreadsheets to help with the sizing.
See for example the VSAN 6.0 Calculator.
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
Yes this is not complete but reasons to take configured is to accommodate peak .
These vms are already been configured and running in fc storage .i think if I add future growth and other overhead following the same logic it can give acceptable results .
However I will compare with what already existing calculator .one of the calculators I checked did not have non uniform workload and also no hybrid options .
Checking what yu provided....
Hi Luc,
for some reasons i am still not able to use calculators .is it possible to build on the script assuming only configured value not the avergae values .
Yes, you can use the configured values for your VMs, but as I remarked earlier, this will not expose over- and under-committed configurations.
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
1:right now this is what i am thinking
in vsan environment the default storage for vms is thin .
however we can make them thick to allocate space right away .
i am doing callculations based on thick disk and max configured values .
so iam trying to put following simple maths in powercli (considering ftt=1 and fault tolerance raid1)
2*total provisioned space+swap+futuregrowth percentage .
all iam stuck how to ensure one esxi failure space .
somehow all these planning is very much explained in vrops but we dont have so iam using powercli .
---------------------------------------------------------------------------------------------------------------------------------
2:also if you could tell me how to find what storage policy configured on vm using powercli
Get-VM | Get-SpbmEntityConfiguration
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
thanks .
lets say by some method we determined that we need 4 TB space(vsan datastore) in 4 esxi hosts .
how to handle one esxi failure in a cluster of 4 so that datastore does not run out of space .
I'm not that experienced in VSAN sizing to give you a simple answer to that.
And I'm not sure if someone can give you a direct answer with the only input being a capacity of 4TB and support for 1 node failure.
It all depends on many factors: the characteristics of your intended load, and the HW you will go for, the intended growth, all-flash or hybrid...
In short, there are no simple rules of thumb afaik.
Why don't you have a go with the vSAN ReadyNode Sizer?
It's for a specific HW type, but at least it might give you some indications of what you should be looking at.
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
it does not have option to go beyond25 vms.
Let me check again .hope yu are login as VMware account not as partner account.
Yes, the My VMware option
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
for some reasons i dont get single site cluster option.only roboand streched clusters.
Don't you see the 'Additional Cluster Information' part?
-----------------------------------------------------------------------------------------------------------------------------------------------
Was it helpful? Let us know by completing this short survey here.
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
oh i thought additional cluster information is for either robo or stereched .
however it seems this is for single site cluster as iam able to click next without choosing roboor streached .
thnaks for your help.