#VCP Datastores are storage units inside vSphere that are formatted in VMFS format and specificially designed for use with vSphere. Traditional Datastores that are backed by a Storage (ISCSI/FC e.g.) are made up of one or more LUNs. ## Extent An extent is backed by a partition on a phyisical device such as a LUN. Usually there will be only one extent, meaning a physical LUN is backing up a single VMFS partition. But this was not always the case, VMFS-3 extent maximum size was 2TB due to the reliance on SCSI-2 addressing and the MBR format. Meaning that if you wanted a datastore bigger than 2TB you had to expand your datastore with multiple extends. VMFS 5 included GPT format and an extent could be as large as 64TB. Generall VMware recommends to avoid using different extents, it gives little to none benefits and adds a much larger management overhead and complexity to your environment. ### Failures When an extent goes offline the whole volume does not need to be affected, depending on the usage of the datastore. If an extent goes offline and there are virtual disk files of a VM on them, they will be unavailable. If the head extent (1st extent of datastore) goes offline it will bring down the datastore because the address resolution resides on them. ### benefits If you have multiple physical devices backing up the datastore, high queue depth can be dealt with more easily in certain scenarios, as they will be sent to different devices. This is debatable though, as you could just bump the queue depth to a single device up instead. ## 🔗Resources - https://blogs.vmware.com/vsphere/2012/02/vmfs-extents-are-they-bad-or-simply-misunderstood.html