Processors and Components: Monitoring and

Monitoring and Maintaining Processors

Recovery

 

In summary, these terms describe the NSAA processor:

Term

Description

Blade

Consists of two Blade Elements (in a duplex system) or three Blade

Complex

Elements (in a triplex system) and up to four logical processors and their

 

associated LSUs. An Integrity NonStop system includes up to four Blade

 

Complexes.

Blade

Consists of a chassis, processor board containing two or four PEs (one

Element

representing each logical processor in the Blade Complex), memory, I/O

 

interface board, midplane, optics adapters, fans, and power supplies.

 

Blade Elements are mounted in a 19-inch computer equipment rack.

Processor

A single Itanium microprocessor with its associated memory. A PE is

element (PE)

capable of executing an individual instruction stream and I/O

 

communication through fiber-optic links.

Logical

One PE from each Blade Element executing a single instruction stream. A

processor

duplex system has two PEs forming a logical processor. A triplex system

 

has three PEs.

Monitoring and Maintaining Processors

To monitor processors, use OSM, the ViewSys product, and other tools. Monitoring and maintaining processors includes:

Monitoring Processors Automatically Using TFDS on page 9-4Monitoring Processor Status Using the OSM Low-Level Link on page 9-5Monitoring Processor Status Using the OSM Service Connection on page 9-5Monitoring Processor Performance Using ViewSys on page 9-7

Monitoring EMS Event Messages on page 4-1

Monitoring Processors Automatically Using TFDS

HP Tandem Failure Data System (TFDS) should be used to proactively monitor processors and manage processor halts. Configured and running before a halt occurs, TFDS can help determine the type of recovery operation needed and:

If TFDS determines that the entire processor should be dumped be reloading, it automatically dumps, then reloads the processor.

If TFDS determines that only the processor element (PE) for one Blade Element needs to be dumped, it reloads the processor, excluding that Blade Element, dumps the Blade Element, then reintegrates it back into the running processor.

Collects the files necessary to analyze the problem.

Sends halt information message to the EMS collector. If configured in OSM, a dial-out message is sent to HP Global Support to notify them of the halt.

For more information on configuring and using TFDS, see the Tandem Failure Data System (TFDS) Manual.

HP Integrity NonStop NS-Series Operations Guide529869-005

9-4

Page 116
Image 116
HP NonStop NS Monitoring and Maintaining Processors, Monitoring Processors Automatically Using Tfds, Term Description