capprofile(1M)

NAME

capprofile — Import, export, display, invalidate and remove Capacity Advisor data for workloads or systems.

SYNOPSIS

Path: /opt/vse/bin

capprofile [-i] [-b begin-time] [-e end-time] [-o] [-y delimiter] [-S ] profileID

capprofile [-l] [ n t v ] [-b begin-time] [-e end-time] [profileID ...]

capprofile [-m i|v] [-b begin-time] [-e end-time] profileID ...

capprofile [-r] profileID

capprofile [-x] [-b begin-time] [-e end-time] [-t] [-y delimiter] profileID ...

DESCRIPTION

The capprofile command manages utilization data collected on the Systems Insight Manager Central Management Server (CMS) for Capacity Advisor. With capprofile, you can export profile data, either as a file formatted for further processing, or as a table with headings. With the capprofile command, you can display the names of systems or workloads that have sufficient utilization data for Capacity Advisor to analyze and, for a particular system or workload, you can display those time intervals that have valid data. The capprofile command also lets you mark utilization data for specified time intervals as invalid.

Refer to the capprofile(4) manpage for information on capprofile import and export file formats.

Options

-b begin-time

Specifies the time to begin using profile data. If omitted, the first available profile time is used.

-e end-time

Specifies the time to stop using profile data. If omitted, the last available profile time is used.

Begin-time and end-time are defined as YYYYMM[DD[hh[mm]]] as follows:

Units	Meaning
YYYY	Year (for example, 2005)
MM	Month of the year (01-12)
DD	Day of the month (01-31)
hh	Hour of the day (00-23)
mm	Minute of the hour (00-59)

If a day, hour or minute field is omitted, its value defaults to zero.

-h

Displays command usage.

-i

Imports utilization data for a workload or system. The capprofile(4) man page describes the format of the text representation of the imported data. The imported text is read from standard-in (stdin). Also see the -o, and -S options.

-l

Lists profile summaries. By default, this is a list of the systems and workloads that have any profile data.

Use -ln to list the profileIDs (the name or unique identification of the profile); use -lt (the default) to list profileIDs and available history; use -lv to list profileIDs, available history, and percent valid.

-m

Marks utilization data as valid (v) or invalid (i) for a specified system or a group of systems or workloads. When a data range is specified, only data within that range will be affected. If no date range is specified, all data is marked. Specify the utilization data to be marked by date range. Marking data by a date range marks every sample in the specified range. Marking utilization data invalid does not remove or alter it; Capacity Advisor just ignores it. Marking utilization data invalid removes all utilization values (CPU, memory, disk and network) from consideration. You cannot, for example, mark only the CPU utilization portion of a sample invalid.

Use -mi to mark data for a specified interval and profileID as invalid; use -mv to mark data as valid.

-o

When used with the -i option, overwrites existing data.

-r profile_ID

Removes all profile data for a specified workload or system from the CMS. This can be used, for example, to release disk space when a workload is retired and its performance data is no longer needed. The profileID names a specific workload. All system name profiles are fully qualified.

-S

Saves the import-compartment attributes. Used in conjunction with -i to import data. Saves the container attributes specified in the import header as attributes for the target profile.

-t

Exports data as a table.

-x

Exports utilization data for the specified profile as a sequence of lines. Each line contains utilization values for a five-minute period. The format is specified in the capprofile(4) manpage. The exported text is sent to standard-out (stdout).

-y delimiter

Specifies the delimiter character used to separate values in imported and exported data. The default delimiter is the comma. Valid delimiters, with or without double quotes, are: comma (,), slash (/), semicolon (;), colon (:) and vertical bar (|); some of these delimiters may need to be quoted or escaped.

EXAMPLES

Export profile data from December 14, 2005 at midnight to midnight December 31, 2005 for the workload with profileID billing3 to the file billing3.txt in the /tmp directory.

capprofile -x -b 20051215 -e 200512312359 billing3 > /tmp/billing3.txt

Import profile data from the file /tmp/billing3.txt for the profile with profileID billing3.

capprofile -i -o billing3 < /tmp/billing3.txt

Mark utilization data as invalid for the profiles of workloads on system prod05 with profileID prod05_wkld. All samples between December 15 2005 at 11 a.m. and December 16 2005 at 12 p.m. (inclusive) are marked invalid.

capprofile -mi -b 2005121511 -e 2005121612 prod05_wkld

List profileID, available history, and percent valid. (Note that a Percent Valid value of less than 100% can occur when there is invalid or uncollected data. For example, there may be one or two uncollected data samples on a node right after it reboots.)

capprofile -lv

Name              Available History                    Percent Valid 
node01.co.com     01/17/06 05:30 pm - 01/19/06 10:10 am     100.00%
node02.com        12/18/05 05:00 pm - 01/20/06 01:05 pm      99.96%
billing3          12/10/05 05:00 pm - 01/20/06 01:05 pm     100.00%
node03.com.OTHER  01/15/06 03:05 pm - 01/18/06 12:55 pm      82.83%
cimserver         01/17/06 05:30 pm - 01/19/06 10:10 am      99.78%

WARNINGS

While the capcollect(1m) command is running, write access to profile data using the capprofile(1M) command (this command) may be delayed.

FILES

/var/opt/vse/profile/bin: Root directory for storing collected data for both capcollect and capprofile. Note that these data files may change without notice from release to release. The directory where these files are stored needs to be sized to hold all the daily utilization profile data, which for each system, is about 24K per day per workload, taking into account minimum block size of 8K. For example, with a CMS monitoring 100 systems, each with an average of 3 workloads, with a need to keep historical data for one year, the file system where /var/opt/vse/profile resides must have space for (100 systems + 300 workloads) * 365 days * 24K = 3.34 Gigabytes. With a file system that allocates storage in blocks greater than 24K, then use that larger block size when calculating storage. Using vseinitconfig(1m) can create a backup of profile data that is more compact.

capprofile(1M)

Technical documentation

» Table of Contents

» Glossary

» Index