2.5 SANtricity Storage Features
SANtricity offers several layers of storage features ranging from security for data at rest, features that
manage host paths, features to manage large-capacity drives that ensure data integrity and efficiently
manage drive faults, and features that provide data protection. The following sections describe many of
the features and provide links to additional information resources.
Drive Encryption
When external key management is enabled from the Settings tile, use the Key Management tab to
generate a CSR file. Use the CSR file on the key management server to generate a client certificate.
Import the client certificate from the Key Management tab to enable secure communication between the
E-Series controllers and the external key management server. For more information about the SANtricity
drive security feature, see the E-Series online help center and TR-4474: NetApp SANtricity Drive
Security: Feature Details Using SANtricity 11.50.
SANtricity Host and Path Management Features
When considering the elements of E-Series multipath functionality, you must understand two concepts.
The first is controller-to-volume ownership and how path failover between controllers is managed through
asymmetrical logical unit access (ALUA). This scenario occurs when the primary paths to an E-Series
volume (I/O paths through the owning controller) are lost. The second concept concerns how the
multipath driver on the host interacts with multiple ports on each E-Series controller (target port group
support, or TPGS) to spread I/O across the interfaces and maximize performance. This section provides a
brief explanation of each concept. For a deep explanation of E-Series multipath behavior, see TR-4604:
Clustered File Systems with E-Series Products: BPG for Media.
The design of the E-Series multipath behavior has evolved from a host multipath driver-managed
scenario (explicit failover) to the new E-Series-led path management model (implicit failover). However,
the E-Series fundamentals have not changed. For example, E-Series systems have asymmetric dual
active controllers for which:
• Volume ownership alternates as volumes are provisioned.
• Write I/O is mirrored to the peer controller.
• Both controllers have access to every volume on the array.
• Both controllers have multiple host ports.
• If one E-Series controller fails, the other controller takes control of all the LUNs and continues to
process I/O.
These attributes allow host multipath drivers to spread I/O across ports on each controller that are
associated to the volumes owned by that controller (TPGS). The drivers use path policies such as least
queue depth and round robin. Depending on the host operating system, the default path policy is one of
these two methods.
When all the paths from a host to one E-Series controller are lost, I/O from that host to the volumes
owned by that controller is routed to ports on the other E-Series controller, which performs “I/O shipping”
across the shelf midplane to the controller that owns the volumes. In parallel, an ALUA timer is set, and
changes in controller-to-volume ownership are delayed until the timer expires. This delay time is long
enough for links to reset and return to service (the default is 5 minutes). After the timer expires, the array
decides whether to initiate a change of volume ownership to the peer controller. The decision is based on
whether the nonowning controller is still receiving more than 75% of the I/O.
Table 4 provides a list of SANtricity host types and the associated support for implicit failover/failback.
To Next Page
Loading...
