Users Guide

NOTE: Consider a VA assigned to server in the lower slot (3 or 4). When a half height server (slot 3 or 4) is replaced
with a full height server, the full height server does not access the VA assigned to lower slots. Inserting a half height
server again, provides access to the VA.
Map or un-map a PERC Virtual Controller to blade:
● Each External Shared PERC 8 card has four virtual adapters (VA). If one or two External Shared PERC 8 cards are
present in the system, then in shared mode you can map or un-map one of the four virtual adapters’.
● If an external PCIE slot is occupied by shared adapter, the virtual adapter mapping can get the current details or
information for VA mapping of the shared storage VA pool.
● Shared device is not supported when the external PCIE slot is occupied by a shared adapter. Using the shared adapter,
you can support shared device by changing the shared storage VA pool.
Fault-Tolerance in Storage Controllers
High Availability (HA) in storage enables availability of multiple integrated components and multiple access points to storage
resources. In case a storage component stops functioning, the server is supported by a second critical component or path
to the available data. High Availability only minimizes downtime by restoring services behind the scenes, in most cases before
the non-functionality is visible, but does not eliminate downtime. Fault Tolerance (FT) makes use of redundant components
within a storage system, which are configured to behave as backup components and are kept in standby mode. Storage
Controllers in fault-tolerant mode prevent disruption of storage services and automatically take over the services of a
component that has stopped functioning. Performance remains consistent throughout this failover process since the redundant
components(controllers) are not used during normal operating conditions.
High Availability with fault-tolerance provides the following benefits:
● Provides uptime for all storage applications even when a controller stops functioning.
● Provides access to critical functions of the chassis at all times.
● Enables server to handle situations when controller stops functioning are becomes faulty.
● Makes use of component redundancy
Using the fault-tolerant feature of controllers, you can manage the tasks associated with shared storage that are achieved
by having an active and passive (peer) controller. Active controller is the controller that is active and monitors all the storage-
related processes. Working status of both controllers are communicated between controllers so that when an active controller
stops functioning, the passive controller acting as a peer hot-spare, takes over in a seamless manner.
NOTE:
CMC displays fault-tolerant data for Shared PERC 8’s with SR-IOV enabled firmware. If a non-SR-IOV card is
attached to the shared storage slots the card does not power on and an alert is generated.
NOTE: Operations such as resetting CMC, which reset the CMC configuration, reset the external fault-tolerant
configuration. As a result, the PERC mode changes to "Safe mode". Disable the fault-tolerance in the external PERC.
Security Key Mismatch
You can create a security key on a controller using an Encryption keyID and a Passphrase. The controller compares only the
Passphrase used while creating security key to identify whether the two controllers have the same security keys. Therefore,
two controllers joining a cluster are fault-tolerant even if they have different Encryption Key ID as long as they have the same
passphrase.
If a security key mismatch is detected between two peer controllers, the fault-tolerant mode changes to ‘Degraded’. A critical
alert is displayed on the Chassis Health page and monitoring may not display proper drive association.
If a security key mismatch is detected, resolve the key mismatch by creating, modifying, or deleting the security key on one of
the controllers, before performing any other storage security operation on the controller. Power cycle the chassis after resolving
the mismatch. Before combining two non-high availability controllers, modify the keys so that they match. This action facilitates
import of secure drives associated with each controller joining the cluster.
For external controllers, modify the keys so that they match before cabling them for fault tolerance. The modification of security
keys facilitates import of secure drives associated with each controller joining the cluster.
178
Managing Chassis Storage