Monitoring WLM

ps

The following ps command has options specific to PRM that WLM uses to define workload groups when dividing resources within a single HP-UX instance:

ps [-P] [-R workload_group]

wlminfo

The wlminfo command, available in /opt/wlm/bin/, displays information about SLOs, metrics, workload groups, virtual partitions or nPartitions, and the current host. To display information about workload groups, specify the group keyword as in the following example. Note that as of WLM A.03.02 you can use the -v option with the wlminfo group command to display gmincpu, gmaxcpu, gminmem, and gmaxmem values, if they are assigned in the active configuration file).

# wlminfo group

Workload Group   PRMID  CPU Shares  CPU Util  Mem Shares  Mem Util  State
OTHERS               1       65.00      0.00        6.00     2.10   ON   
g2                   2       15.00      0.00       64.00    32.43   ON   
g3                   3       20.00      0.00       30.00     9.17   ON   

To display information about partitions, specify the par keyword. In the following example, the ‘Intended Cores’ column shows the number of CPU resources (cores) that WLM wants to allocate to the partition, while the ‘Cores’ column shows the current number of active cores. The number of intended and active cores is usually the same except when WLM is in the process of modifying a partition or is operating in passive mode. (In passive mode, the intended core allocation is not made; the partition retains the current number of active cores.) The ‘Cores Used’ column shows the CPU (core) utilization of the partition. The ‘Interval’ column shows the WLM interval—the frequency at which WLM checks for new performance data for the workload and then adjusts core allocations.

# /opt/wlm/bin/wlminfo par
 
Hostname      Intended Cores  Cores  Cores Used  Interval   
north                    2      2        1.3        6
south                    3      3        2.1        6
east                     1      1        0.4        6
west                     2      2        1.7        6
northwest	                3      3        2.3        6
northeast                2      2        1.4        6

The wlminfo host command displays information pertaining to the local host (default) or a specified host, including the number of CPU resources (cores) on the host and the number being used as well as the WLM interval, as in the following example (for local host west):

# /opt/wlm/bin/wlminfo host
 
Hostname       Cores  Cores Used  Interval   
localhost          2         1.7         6

For more information on the use of the wlminfo command, see Appendix A “WLM command reference” and wlminfo(1M).

wlmgui

The wlmgui command, available in /opt/wlm/bin/, graphically displays information about SLOs, metrics, workload groups, partitions, and the current host.

prmmonitor

The prmmonitor command, available in /opt/prm/bin/, displays current configuration and resource usage information, as in the following example. The ‘CPU Entitle’ column indicates the CPU entitlement (allocations) for the group. The ‘Upper Bound’ column indicates the per-group consumption cap; the column is blank for each group because the CPU consumption cap is not available with WLM. The ‘LCPU State’ column indicates the Hyper-Threading setting (ON or OFF) for PSET-based groups. The column is blank if the system does not support Hyper-Threading or if a group is not a PSET-based group.

PRM configured from file:  /var/opt/wlm/tmp/wmprmBAAa06335
File last modified:        Tue Sep 26 08:35:23 2006 
 
HP-UX vpar02 B.11.31 U ia64    09/26/06
 
Tue Sep 26 08:43:16 2006    Sample:  1 second
CPU scheduler state:  Enabled
                                   CPU  Upper    CPU    LCPU
PRM Group               PRMID  Entitle  Bound   Used   State
 ____________________________________________________________
(PRM_SYS)                   0                  0.00%
OTHERS                      1   12.50%         0.00%
grp2                    65536   12.50%         0.00%     OFF
grp3                   131072   25.00%         0.00%      ON
grp4                   196608   50.00%         0.00       ON
PRM application manager state:  Enabled  (polling interval: 30 seconds)

prmlist

The prmlist command, available in /opt/prm/bin, displays current CPU allocations plus user and application configuration information. The ‘Upper Bound’ column indicates the per-group consumption cap; the column is blank for each group because the CPU consumption cap is not available with WLM. The ‘LCPU Attr’ column indicates the Hyper-Threading setting (ON or OFF) for a PSET-based group; the column is blank if the system does not support Hyper-Threading or the group is not a PSET-based group.

PRM configured from file:  /var/opt/wlm/tmp/wmprmBAAa06228
File last modified:        Thu Aug 24 08:35:23 2006 
 
PRM Group                  PRMID               CPU    Upper    LCPU
                                       Entitlement    Bound    Attr
-------------------------------------------------------------------
OTHERS                            1       65.00%
g2                                2       15.00%
g3                                3       20.00%
 
PRM User                       Initial Group                  Alternate Group(s)
--------------------------------------------------------------------------------
root                           (PRM_SYS)                     
 
PRM Application                Assigned Group                 Alternate Name(s)
--------------------------------------------------------------------------------
/opt/perl/bin/perl             g2                             loop2.pl          
/opt/perl/bin/perl             g3                             loop3.pl          

GlancePlus

The optional HP product GlancePlus allows you to display resource allocations for the workload groups as well as list processes for individual workload groups.

wlm_watch.cgi

This CGI script, available in /opt/wlm/toolkits/apache/bin/, allows you to monitor WLM using a web browser interface to prmmonitor, prmlist, and other monitoring tools.

For information on setting up this script, see wlm_watch(1M).

Status and message logs

WLM provides the following logs:

For information on these logs, including sample output, see the section “Status information WLM provides”.

Event Monitoring Service (EMS)

EMS (Event Monitoring Service) polls various system resources and sends messages when events occur. The WLM wlmemsmon command provides numerous resources for event monitoring. For information on these resources, see wlmemsmon(1M).