wlminfo(1M)wlminfo(1M)NAMEwlminfo - HP-UX Workload Manager (WLM) information monitor
SYNOPSIS
[cmd]
[-l] [-o] [-v] [-b { 0 | 1 }] [-q] [-c]
-s slo1 [-s slo2 ...]
[-l] [-o] [-v] [-b { 0 | 1 }] [-q] [-c]
-g grp1 [-g grp2 ...]
[-l] [-o] [-v] [-b { 0 | 1 }] [-q] [-c]
-m met1 [-m met2 ...]
[-l] [-o] [-v] [-b { 0 | 1 }] [-q] [-c]
[-l] [-o] [-b { 0 | 1 }] [-q]
[-m met1] [-m met2 ...]
[-l] [-o] [-b { 0 | 1 }] [-q]
[-l] [-o] [-b { 0 | 1 }] [-q] [-S] [-v]
-g grp1 [-g grp2 ...]
[-l] [-o] [-b { 0 | 1 }] [-q] [-S] [-v]
[-l] [-o] [-b { 0 | 1 }] [-q]
[-l] [-o] [-b { 0 | 1 }] [-q]
-h host1 [-h host2 ...]
[-l] [-o] [-b { 0 | 1 }] [-q]
[-l] [-t secs] [-n cnt] [-p] [-v] [-b { 0 | 1 }]
-g grp1 [-g grp2 ...]
[-l] [-t secs] [-n cnt] [-p] [-v] [-b { 0 | 1 }]
DESCRIPTION
The utility provides various WLM data. You indicate the type of data to
display by specifying a command with Commands include and Each command
has its own options.
OPTIONS
Displays usage information and exits. If you specify
cmd, the usage information is limited to cmd data. This option
overrides all other options and commands.
Displays version information and exits. This option overrides all com‐
mands
and any options other than
Launches
which displays various data in a graphical user interface (GUI).
This option overrides all commands.
Be sure WLM is enabled and your environment variable is set before
you use
Launching in this mode requires Java(tm) Runtime Environment ver‐
sion 1.5 or later in /opt/java1.5/jre/bin/java, and for PRM-based
configurations, PRM C.03.00 or later. (To take advantage of the
latest updates to WLM and the GUI, use the latest version of PRM
available.)
For more information on this interface, see its online help.
COMMANDS
slo [options]
Displays data about SLOs, including their priorities and associated
workload groups. Options to this command are:
Displays data only for the specified SLOs. (You can use at most one
type of option in the command line: or However, you can use multi‐
ple instances of an option.)
Displays SLOs only for the specified workload groups. (You can use at
most one type of option in the command line: or However, you can
use multiple instances of an option.)
Displays only the SLOs based on the specified metrics. These SLOs have
either a statement or a statement that includes met. (You can use
at most one type of option in the command line: or However, you
can use multiple instances of an option.)
Performs live updates of the data, updating data once every WLM inter‐
val.
Consider using the option to limit the printing of column banners
when you use
Displays in verbose mode, which includes goal and metric values where
applicable. These columns do not apply to all SLOs. Also, data
appears in these columns only if WLM received it in the current
interval. To see more complete metric data, use the command
Displays SLO data from the file /var/opt/wlm/wlmdstats. (You must use
the
option to to create the wlmdstats file when you activate your con‐
figuration.)
This data may contain information on SLOs from multiple (previous)
WLM configurations. It is not separated based on the configuration
from which it came.
Using presents data from multiple configurations; however, using
different logging options with namely the various combinations of
and will produce variations in the data displayed.
If the wlmdstats file is up-to-date, using both and will cause
data from the last WLM interval to appear twice.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using either the or options.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
Quiets (suppresses) the message updates from /var/opt/wlm/msglog
that follow the "New messages:" text in output.
Sorts SLOs first by workload group, then priority, followed by state.
metric [options]
Displays data about metrics, including their values. Options to this
command are:
Displays data only for the specified metrics.
Performs live updates of the data, updating data once every WLM inter‐
val.
Consider using the option to limit the printing of column banners
when you use
Displays metric data from the file /var/opt/wlm/wlmdstats. (You must
use
the option to to create the wlmdstats file when you activate your
configuration.)
This data may contain information on metrics from multiple (previ‐
ous) WLM configurations. It is not separated based on the configu‐
ration from which it came.
Using presents data from multiple configurations; however, using
different logging options with namely the various combinations of
and will produce variations in the data displayed.
If the wlmdstats file is up-to-date, using both and will cause
data from the last WLM interval to appear twice.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using either the or options.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
Quiets (suppresses) the message updates from /var/opt/wlm/msglog
that follow the "New messages:" text in output.
group [options]
Displays data about workload groups, including their CPU and memory
shares (if applicable). If the configuration data does not include
workload groups (the data is only for hosts), produces no output.
Options to this command are:
Displays data only for the specified workload groups.
Performs live updates of the data, updating data once every WLM inter‐
val.
Consider using the option to limit the printing of column banners
when you use
Displays group data from the file /var/opt/wlm/wlmdstats. (You must use
the
option to to create the wlmdstats file when you activate your con‐
figuration.)
This data may contain information on workload groups from multiple
(previous) WLM configurations. It is not separated based on the
configuration from which it came.
If the wlmdstats file is up-to-date, using both and will cause
data from the last WLM interval to appear twice.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using either the or options.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
Quiets (suppresses) the message updates from /var/opt/wlm/msglog
that follow the "New messages:" text in output.
Shows statistics for the
group. (This group is for system processes.)
Includes additional columns to display the group's
and values. These columns are omitted when live data is not being
displayed (for example, when is specified).
host [options]
Displays data about the host. Options to this command are:
Performs live updates of the data, updating data once every WLM inter‐
val.
Consider using the option to limit the printing of column banners
when you use
Displays host data from the file /var/opt/wlm/wlmdstats. (You must use
the
option to to create the wlmdstats file when you activate your con‐
figuration.)
If the wlmdstats file is up-to-date, using both and will cause
data from the last WLM interval to appear twice.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using either the or options.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
Quiets (suppresses) the message updates from /var/opt/wlm/msglog
that follow the "New messages:" text in output.
par [options]
Displays data about virtual partitions or nPartitions--if is running.
Options to this command are:
Displays data only for the specified hosts.
Performs live updates of the data, updating data once every WLM inter‐
val.
Consider using the option to limit the printing of column banners
when you use
Displays virtual partition or nPartition data from the file
/var/opt/wlm/wlmpardstats. (You must use the option to to create
the wlmpardstats file when you activate your global arbiter con‐
figuration.)
If the wlmpardstats file is up-to-date, using both and will cause
data from the last global arbiter interval to appear twice.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using either the or options.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
Quiets (suppresses) the message updates from /var/opt/wlm/msglog
that follow the "New messages:" text in output.
proc [options]
Displays data about the most active processes. Options to this command
are:
Displays data only for the specified workload groups.
Performs live updates of the data, updating data once every sample
period.
Consider using the option to limit the printing of column banners
when you use
Sets the duration of the sample period to secs seconds. By default,
the duration is two seconds. Output is displayed at the end of the
sample period.
Sets the number of processes to show in the output to the cnt most
active processes. By default, 15 processes are shown.
Displays process lineage.
Displays verbose output, revealing more of the process command line.
Prints the column banner 0 or 1 times.
Use this option to reduce the printing of the column banner when
using the option.
For example, when using the option to display live updates, the
banner is printed every update. Using the banner is printed only
when you first start each update is printed without a banner.
COLUMN DESCRIPTIONS AND OTHER OUTPUT
The output from varies with the command you use. For information on a
certain command's output, see the corresponding section below.
Other output from includes:
New messages
If any messages are written to /var/opt/wlm/msglog while is run‐
ning, the messages will be shown in the output preceded by "New
messages:". You can quiet (suppress) this behavior by including
the option.
date/time
Data being displayed is from the interval ending with the date
and time shown in the banner.
*PASSIVE*
If WLM or its global arbiter is running in passive mode, which
is enabled with the option to and you will see "*PASSIVE*" in
the banner.
Using passive mode (by including the option with removes the
feedback loop that WLM uses in attempting to meet SLOs. For
information on the effects of a missing feedback loop--and gen‐
eral limitations of passive mode, see the section PASSIVE MODE
VERSUS ACTUAL WLM MANAGEMENT in the wlm(5) manpage.
wlminfo slo
When you enter the column headers are:
SLO Name Group Pri Req Shares State Concern
When you use the option to the command, you get two additional column
headers, Metric and Goal:
SLO Name Group Pri Metric Goal Req Shares State Concern
These column headers are described below.
SLO Name The name of an SLO in the current configuration
Group The workload group corresponding to the SLO (the group
for which the SLO was created)
A ´-´ appears in this column if the data is for the
host rather than a group based on a PSET or FSS group.
You may see a group named in this column if you have
set in your WLM configuration. For more information,
see the wlmconf(4) manpage.
You may see a group named in this column if data is
not available for the interval.
Pri The priority of the SLO
Metric (Printed only when you use the option) The value of
the metric indicating the SLO's performance
A ´?´ appears in this column if no metric was received
for the interval, resulting in the metric being
unknown.
Goal (Printed only when you use the option) The value of
the goal: The SLO passes depending on how the value in
the Metric column compares to this value
A ´-´ appears in this column if the SLO does not have
a statement.
Req The CPU shares request made by the SLO's controller
A ´-´ appears in this column if the SLO is OFF/Dis‐
abled.
Each SLO has a controller making CPU requests on its
behalf; WLM considers these requests, in conjunction
with their priorities, to determine CPU allocations.
Shares The CPU shares allocated to the SLO's corresponding
workload group
A ´?´ appears in this column if data is not available
for the interval.
State An SLO can have one of five states:
OFF The SLO is disabled, due to a or statement
INIT has not yet received a metric for the SLO, so
determining whether it is passing or failing is
not possible
DEAD The data collector that generates the metric
used in a statement or statement in the SLO is
no longer running
FAIL WLM's daemon considers the SLO's goal to be
unsatisfied
PASS WLM's daemon considers the goal to be satisfied
Concern
This column highlights information that could possibly
indicate configuration issues or other issues that
require attention. The concerns for an SLO are:
Disabled The SLO is disabled due to a or statement
Metric There is a problem with the metric, such as
not receiving an initial value
Child A child process, typically a data collec‐
tor, has died
Priority The SLO's request was not satisfied because
there were not enough CPU resources remain‐
ing after satisfying a request from a
higher priority or the same priority
Max The SLO's request was not satisfied, and in
fact was reduced, because it exceeds a max‐
imum specified in the WLM configuration,
such as a value or a value
wlminfo metric
When you enter the column headers are:
Metric Name PID State Value
These column headers are described below.
Metric Name A metric in the WLM configuration
Metric names that begin with an underscore (_) are
reserved
PID The process ID of the data collector that provides
the metric to the WLM daemon,
State A metric can have one of four states:
INIT No initial value has been received for the
metric--it has not been updated since the
configuration was activated
DEAD The data collector corresponding to the met‐
ric is not running
OLD The metric's value was not updated in the
interval
NEW The metric was updated with a new value in
the interval
Value The metric's value
wlminfo group
If the configuration data does not include workload groups (the data is
only for hosts), produces no output.
When you enter the column headers are:
Workload Group PRMID CPU Shares CPU Util Mem Shares State
When you use the option to the command, you get four additional column
headers--CPU Min, CPU Max, Mem Min, and Mem Max. Note that with these
additional columns, the width of the display exceeds 80 characters.
When you enter the column headers are:
Workload Group PRMID CPU Shares CPU Min CPU Max CPU Util Mem Shares
Mem Min Mem Max Mem Util State
These column headers are described below.
Workload Group
The name of a workload group
A ´-´ appears in this column if the data is for the
host rather than a group based on a PSET or FSS
group.
You may see a group named in this column if you have
set in your WLM configuration. For more information,
see the wlmconf(4) manpage.
PRMID The group's PRM ID
CPU Shares The amount of CPU resources (cores) available to the
group for the current interval; see NOTE below. (A
core is the actual data-processing engine within a
processor, where a processor might have multiple
cores.)
CPU Min The group's value
A ´-´ appears in this column if the value is not
assigned in the current configuration.
CPU Max The group's value
A ´-´ appears in this column if the value is not
assigned in the current configuration.
CPU Util The amount of CPU resources (cores) the group used
in the previous interval (See NOTE below.)
Mem Shares The amount of memory available to the group for the
current interval
A ´-´ appears in this column if memory management is
not being used.
Mem Min The group's value
A ´-´ appears in this column if the value is not
assigned in the current configuration.
Mem Max The group's value
A ´-´ appears in this column if the value is not
assigned in the current configuration.
Mem Util The amount of memory the group used in the previous
interval (See NOTE below.)
State A group can have one of two states:
ON Normal
OFF The group is a transient group and has only
disabled SLOs; consequently, it has been
removed
It will return when it has at least one active
SLO
The "CPU Util" and "Mem Util" values represent the amount of CPU
or memory resources the group used in the previous WLM interval.
However, the group's "CPU Shares" and "Mem Shares" values repre‐
sent the amount of CPU or memory resources from the current WLM
interval. As a result, comparisons of the numbers could lead
mistakenly to the conclusion that the usage exceeds the shares.
wlminfo host
When you enter the column headers are:
Hostname Cores Cores Used Interval
These column headers are described below.
Hostname The name of the local host
Cores The number of CPU resources (cores) on the local host
Cores Used The CPU utilization on the local host
Interval The length (in seconds) of the WLM interval on the
local host
wlminfo par
If is running, when you enter the column headers are:
Hostname Intended Cores Cores Cores Used Interval
These column headers are described below.
Hostname The name of a virtual partition or an nPartition
Intended Cores
The number of CPU resources (cores) that WLM wants to
give to the host. Typically, this value and the
value in the Cores column differ only when WLM is in
the process of modifying the partitions or when you
are using passive mode. (In passive mode, WLM's
intended core allocations are never made: Partitions
retain their current number of active cores.)
Cores The number of active CPU resources (cores) in the
virtual partition or nPartition
Cores Used The CPU utilization on the local host
Interval The length (in seconds) of the WLM interval on that
virtual partition or nPartition
The "Cores Used" value represents the amount of cores the work‐
load used in the previous global arbiter interval. However, the
"Intended Cores" and "Cores" values represent the current state
of the partitions. As a result, comparisons of the numbers
could lead mistakenly to the conclusion that a workload's CPU
resource usage exceeds the number of CPU resources allocated to
the workload--or that the workload's CPU resource usage exceeds
its intended allocation.
wlminfo proc
When you enter the column headers are:
Workload Group PRMID PID TTime Time Command
When you use the option to the command, you get more of the process
command line.
These column headers are described below.
Workload Group
The name of a workload group (for workload groups
based on FSS groups or PSETs) or a ´-´ if the data
is for a host.
PRMID The group's PRM ID (for workload groups based on FSS
groups or PSETs) or a ´-´ if the data is for a
host.
PID The ID for the process
TTime Total CPU time in seconds consumed by the process
since startup
Time Total CPU time in seconds consumed by the process
during the sample period (The output is sorted on
this column.)
Command The name of the process
DEPENDENCIES
Using the option requires you to initiate logging (with the option)
when you activate your WLM configuration with or your WLM global
arbiter configuration with
NOTES
To ensure the smooth running of the operating system and key HP-UX dae‐
mons, WLM runs these system processes in a special workload group
called This group is not restrained by WLM: It consumes system
resources as needed. As a result, a workload group's CPU usage (shown
in the "CPU Util" column) may be less than its CPU shares because
PRM_SYS requires some of the group's resources. However, the low usage
could also be the result of the group's low CPU demands, or a combina‐
tion of the two factors. Also, CPU usage might at times be slightly
above the allocation due to dynamics in the CPU scheduler that WLM
uses. Likewise, if memory management is enabled, a workload's memory
usage may be less than the number of memory shares for reasons similar
to those just stated. It could also be slightly above the memory shares
value due to extreme paging pressure or when the current group alloca‐
tion is being reduced.
AUTHOR
was developed by HP.
FEEDBACK
If you would like to comment on the current HP-UX WLM functionality or
make suggestions for future releases, please send email to:
wlmfeedback@rsn.hp.com
SEE ALSOwlmd(1M), wlmpard(1M), wlm(5)
HP-UX Workload Manager User's Guide (/opt/wlm/share/doc/WLMug.pdf)
HP-UX Workload Manager homepage (http://www.hp.com/go/wlm)
wlminfo(1M)