beautypg.com

Process statistics – HP Neoview Release 2.4 Software User Manual

Page 45

background image

Table 3-7 NODE_STATS_V1 Field Definitions (continued)

Description

Data Type

Field Name

Time, in microseconds, that the
processing node spent executing
interrupt handlers.

LARGEINT

INTR_BUSY_TIME

Number of I/O disk transfers
performed by disk processes in
this processing node.

LARGEINT

DISC_IOS

Number of times the required
block was found in cache during
an I/O operation.

LARGEINT

CACHE_HITS

Number of free memory frames
(physical pages) when the
processing node was loaded.

LARGEINT

STARTING_FREE_MEM

Number of free memory frames
(physical pages) at the end of the
measurement interval.

LARGEINT

ENDING_FREE_MEM

Number of bytes read by this
processing node, if NODE_ID is
1. Othewise, this field is not
populated.

LARGEINT

SVNET_READ_BYTES1

Number of bytes written by this
processing node, if NODE_ID is
1. Othewise, this field is not
populated.

LARGEINT

SVNET_WRITE_BYTES1

Number of bytes read by this
processing node, if NODE_ID is
16. Othewise, this field is not
populated.

LARGEINT

SVNET_READ_BYTES16

Number of bytes written by this
processing node, if NODE_ID is
16. Othewise, this field is not
populated.

LARGEINT

SVNET_WRITE_BYTES16

Process Statistics

The two views used for access to process statistics differ in that one includes a record for each
measured process, whereas the other aggregates information about measured processes of specific
types. If process data collection is enabled, both views are populated with data.

NOTE:

The process statistics are not turned on by default and are scheduled to be deprecated.

If you want HP Support to enable it, please contact your HP Support representative.

The measurements apply to four key process types:

The Neoview SQL compiler (MXCMP)

The Neoview SQL executor server process (MXESP)

The Neoview NDCS server (MXOSRVR)

The Repository collector processes (MMCOLLEC)

By default, a measurement of each of these processes occurs every five minutes. Aggregates are
computed every five minutes with a 15–minute delay: that is, aggregate data becomes available
from the Repository 15 minutes after the end of the sampling interval.

Process Statistics

45