HP-UX vPars and Integrity VM V6.3 Administrator Guide

taken for an I/O card replacement in a vPars and Integrity VM environment vice versa an SD
server configured as native HP-UX.
11.4.1 CRA on a VSP
On a standalone SD2 server, before an online replacement of an I/O card, a CRA of all the system
resources that are impacted by the unavailability of the card in question is performed. Only when
this analysis indicates that there is no impact to any of the critical system resources and the operation
is safe and will not cause disruption in the functioning of the system, the olrad command proceeds
to prepare the I/O card for replacement.
With vPars and Integrity VM V6.3, when the SD2 server is configured as a VSP, then, in addition
to performing a CRA on the VSP, the olrad command triggers a similar CRA on all active vPars
or VM guests that have IO resources backed by the card being replaced. Only when the CRA
succeeds across all the active guests and the VSP, does the olrad command proceed to prepare
the I/O card for replacement. HP recommends that the VSP administrator runs the olrad command
with the -C option to check the criticality of the I/O card across the VSP and all active guests
before attempting to perform an online replacement operation.
This command lists one of the following severity levels:
No affected resources in use either on the VSP or on any
of the active vPars or VM guests.
CRA_SUCCESS
Resources are in use on affected devices either on the VSP
or on any of the active guests, but none are deemed
critical.
CRA_WARNING
Resources are in use on affected devices either on the VSP
or on any of the active guests, and there is a probable
data loss. The operation must only proceed with the
permission of the user.
CRA_DATA_CRITICAL
Resources are in use on affected devices either on the VSP
or on any of the active guests, and the operation is likely
to bring down the VSP or one or more of the active vPars
or VM guests.
CRA_SYS_CRITICAL
Indicates that an internal CRA error was encountered and
the CRA across the VSP or one or more of the active vPars
or VM guests cannot be completed.
CRA_FAILURE
For more information about CRA framework, results and reports generated by CRA for various
configurations, and scenarios on a system installed with the HP-UX 11i v3 operating system, see
the white paper Critical Resource Analysis.
11.4.2 NPIV devices
Starting with vPars and Integrity VM V6.3, an FC card in an OLR capable PCI slot can be replaced
without having to bring down active vPars or VM guests that are configured with NPIV HBAs
backed by the FC card. This can be done as long as the NPIV devices impacted are not critical
for the operation of any of the vPar or VM guest. In addition, there must be no impacted devices
on the VSP being critical to the operation of the VSP.
When a CRA check is done on an FC card seen by the VSP, the olrad command triggers a
parallel CRA check on each of the vPar or VM guests with NPIV HBAs backed by the FC card.
These CRA checks in the vPar or VM guests take into account all aspects that are considered by
the Mass Storage CRA checks on a native SD2 server. This includes scenarios like boot and alternate
boot path configurations, swap and dump device configurations, Serviceguard lock disk
configurations, File system and Volume manager configurations, I/O in progress configuration,
and various SAN storage configurations. Only when all these aspects are analyzed per vPar or
164 PCI OLR support on VSPs