Hi,
1) A single VMware Paravirtual SCSI Controller 0 for all VM's irrespective if they have 1x disk or more (generally up to 3x or 4x vdisks)
VS
2) A VMware Paravirtual SCSI Controller per vdisk (Up to 4x SCSI Controllers)
Is there any issue with Option 1?
What are the performance limits of having a single VMware Paravirtual SCSI Controller ?
Would you always create a new VMware Paravirtual SCSI Controller when adding a new vdisk to a VM or this is for extreme case where there is a certain I/O, examples?
Queue depth: https://kb.vmware.com/s/article/2053145
Coalescing averaging: Interrupt coalescing rate determination is on a per virtual controller basis, e.g. for some DB parts (transaction log etc.) you want to have synchronous IO (latency sensitive) issued and completed as fast as possible, if you also have a data disk on that controller the flood of async IO would widen the interrupt window for the serial IO. Again, mostly relevant in performance critical / partially latency sensitive workloads.
ok so looks to be a requirement only for heavy I/O Applications and guess you can easily add an additional SCSI Controller and change a specific disk/s to connect to it (Think it would just be a matter of shutting down VM and doing the changes on vSphere, no impact on VM and should just boot fine after change)
