The purpose of slack space in vSAN is to allow for maintenance and failure operations. This includes tasks such as data evacuation during maintenance mode, resynchronization after a service interruption, application of storage policy changes, and standard background activities like balancing data across drives. Keeping adequate slack space ensures that these operations can be performed without running out of storage capacity.
When one host loses connectivity in a 4-node vSAN cluster with RAID-5 policy, the object remains in a reduced availability state but does not undergo a rebuild. RAID-5 requires a minimum of four hosts to ensure data protection and availability. With the loss of one host, there are insufficient resources to perform a rebuild, so the object remains accessible but does not re-protect itself, leading to reduced availability without a rebuild.
When a vSAN 5-node cluster with two hosts fails permanently, the following are the potential outcomes: For Group A VMs with a storage policy of FTT=0 (Failures to Tolerate), any host failure would result in data inaccessibility since there is no redundancy. Therefore, the VMs in Group A might experience data inaccessibility. For Group B VMs with a storage policy of FTT=2, the system allows two hosts to fail and still keep the data accessible. However, if a third host fails, the data would become inaccessible due to the lack of sufficient redundancy to maintain quorum. Hence, the VMs in Group B will become inaccessible if a third host fails, as the system cannot rebuild or distribute the data to the remaining hosts.
When a drive is physically removed from a vSAN node, the system marks all components on that device as absent. This is in order to handle transient issues without causing unnecessary rebuilds. Additionally, vSAN waits for the configured delay timer before it starts the rebuild process. This delay helps to ensure that temporary issues do not trigger a rebuild, which would otherwise consume considerable resources.
To increase the vSAN datastore storage capacity, two effective methods are to add a host to the cluster and to add a disk group to each host. Adding a host increases the overall resources available to the cluster, including storage capacity. Adding a disk group to each host directly increases the number of capacity devices, thereby increasing the total storage capacity of the vSAN datastore.