NonStop Operations Guide for H-Series and J-Series RVUs

During a power failure, a ServerNet/DA remains operational during the power-fail delay time,
but the external modular disk and tape subsystems attached to it do not. This type of situation could
result in data-integrity problems if the system software continues processing data from an external
disk drive or tape drive during a short power outage.
If a power failure occurs and the processors resume operations but one or more external devices
fail, data integrity problems can occur. The application programs must be resilient to such device
failures.
ESS Cabinets
ESS cabinets require a site UPS. ESS cabinets are not powered off programatically with a power
failure because they may be attached to other systems.
Networking CLIMs
During a power failure, Networking CLIMs will not be put into a low-power mode. Shutdown of
the CLIM may not be graceful. Refer to the Cluster I/O Protocols (CIP) Configuration and
Management Manual for more information about how to shut off the CLIM.
Storage CLIMs
During a power failure, a Storage CLIM without Write Cache Enable will not be put into low-power
mode. Shutdown of the CLIM may not be graceful. External modular disk and tape subsystems
attached to the CLIM do not maintain power. This type of situation could result in data-integrity
problems if the system software continues processing data from an external disk drive or tape drive
during a short power outage.
A Storage CLIM with the Write Cache Enable option selected and a local UPS on that rack can
preserve power long enough for the cached data to be flushed to disk. See the SCF Reference
Manual for the Storage Subsystem for information on setting the WRITECACHE disk attribute to
ENABLED (the default is DISABLED).
SAS Disk Drive Enclosures
SAS disk drive enclosures respond to a power failure as described in “NonStop Cabinets (Modular
Cabinets) (page 171). If a disk volume is configured with the WRITECACHE disk attribute enabled
to allow for faster writes to SAS disks, HP recommends that the volume be connected to a UPS to
give the drives enough time to flush the write cache in the event of power loss.
All disk drives in SAS enclosures connected to G6 CLIMs, including any partitions they might have,
are supported by OSM. OSM will create an alarm and send a problem incident report when a
SAS solid state drive (SSD) approaches or has reached its wear-out threshold.
For more information, refer to the SCF Reference Manual for the Storage Subsystem.
Air Conditioning
Unless the site UPS provides for maintaining air conditioning, it is possible that the temperature
could rise in the computer room complicating the ability of a system to survive overheating before
shutting itself down. For information about site overheating problems, refer to the Planning Guide
for your NonStop system.
Preparing for Power Failure
To prepare for power failures, set ride-through time, configure OSM power fail support, and
regularly monitor power supplies and batteries.
172 Power Failures: Preparation and Recovery