Archive for the ‘SiteScope’ Category

Integrating SiteScope 11.30 Events with BSM 9.25

  1. Install SiteScope 11.30
  2. Extract "Software_HP_Operations_Agent_v11.14_Windows_ISO_TC097-15046" iso locally on sis server
  3. CD to that extracted dir (run as admin)
  4. Run "cscript oainstall.vbs -i -a" and click ok to any prompts, etc.
  5. Select Start > All Programs > HP SiteScope > Configuration Tool (run as admin)
  6. In the "Configure HP Operations Agent installed separately" option select "Configure HP Operations Agent" to complete the installation of the HP Operations agent.
  7. Restart SiteScope server.
  8. In SIS UI, select Preferences > Integration Preferences:

    • Click the "New Integration" button and select "HP Operations Manager Integration".
    • Configure the OM/BSM path (in distributed environment, point to gateway server)
    • Click "Connect" to connect the agent to the BSM server. This sends a connection request from the agent to the specified server.
    • Leave this window open
  9. Accept the agent connection request on the BSM Gateway/Web Processing server

    • In BSM 9.20 or later UI, go to Admin > Operations Manager > Certificate Request, locate and grant your certificate request.
  10. Install the log policies on the SiteScope server

    • After the certificate request has been granted on the BSM server, back in SIS UI, click "Install Policies" in the HP Operations Manager Integrations Main Settings panel of the HP Operations Manager Integration dialog box. This installs and signs the preconfigured log file policy file on the HP Operations agent.
    • Leave this window open
  11. Check connection status and send test message from the SiteScope server – optional

    • In the HP Operations Manager Integration dialog box, expand the HP Operations Manager Integrations Main Settings panel, and click "Analyze".
    • To check that the agent is connected to the BSM server, type a message in the Test message text box, and click "Send Test Message".
    • If the test is successful, the text message is displayed in the Operations Management Event Browser in BSM. This message is generated by a basic opcmsg policy command (i.e. opcmsg a=a msg_grp=test o=o msg_t=test).
  12. Enable SiteScope to send events to HPOM or Operations Management

    • In the HP Operations Manager Integration dialog box, expand the HP Operations Manager Integrations Main Settings panel, and in the Configuration Settings area select "Enable sending events".
    • Leave all other options at default
    • Save and close the window – done.
  13. Note: LWSSO will automatically be configured between BSM and SIS.
 
Reference: SIS_1130_OM_Integration_Guide.pdf (pages 21 – 32)

Integrating Sitescope to HP Operations Manager (BSM)

  1. Stop Sitescope Service on host
  2. Launch Configuration Tool from Start Menu (run as administrator)
  3. Install the Operations Agent
  4. On the SiteScope server (where the OM agent is running):

    1. Execute "ovconfget sec.core CORE_ID" from command line. Copy the results.
  5. On the BSM DPS machine: (<coreid> is output of step A)

    1. Execute "ovcert -exporttrusted -file \dps" from command line
    2. Execute "ovcm -issue -file \sitescope00 -name:  <sitescope_server_FQDN> -coreid <coreid>" from command line
    3. Copy the created files from DPS to SiteScope server
  6. On SiteScope server:

    1. Execute "ovcert -importtrusted -file \dps" from command line.
    2. Execute "ovcert -importcert -file \sitescope00" from command line.

Note: One reason you might perform this kind of manual integration is if port 383 is not open between the SiS and DPS servers.

Custom RHEL Memory Monitor – Part II

In the last post I described how to create a custom memory monitor using a Script Monitor.  The Script Monitor, while functional, may not be desirable as the "Value" returned does not fit into default Threshold/Alert configurations.  Which basically means the monitor can run, but cannot be Alerted upon.

Instead, by editing RedHatEnterpriseLinux.config, located on the SiS host (E:\SiteScope\templates.os dir), we can define the command being executed to determine the REAL "Physical memory used %" for RHEL systems.

Solution:

  • Open the RedHatEnterpriseLinux.config file, locate the following:

#
id=physicalMemory
command=/usr/bin/free -b
total=2
free=4
matchLine=Mem:
#

  • Replace that with the following:

#
id=physicalMemory
command=echo `free | grep - | awk '{ print $4 }'`" "`free | grep Mem
|awk '{ print $2 }'`
total=2
free=1
#

  • Restart SiteScope Server

This monitor will now show the correct Physical memory used % for RHEL systems.

Custom SiteScope Memory Monitor – Part 1

In certain environments, a custom memory monitor may be required to monitor Unix based hosts.  Why? A Unix System Admin might describe it as follows:

Unix/Linux will fill any free memory with disk buffers and caching, but that should not be counted as “used” memory, because if an application needs it, the kernel will give it up. The proper way to calculate memory usage is to pull the total used and subtract whatever amount of cache and buffers are being used.

echo 'free | grep - | awk '{ print $4 }''/'free | grep Mem |awk '{print $2 }'' *100 | bc -l

On the SiteScope host, create a 'CustomMemoryMonitor.sh' file in /SiteScope_Root/remote.script/ directory. This can be done using Notepad.

Insert the script above into your .sh file and save it.

In the SiteScope interface, create a Script Monitor directed at your Unix host with the following settings:

  • Script: USE COMMAND
  • Match expressions: /[0-9]{2}.[0-9]{2}/
  • Remote script Command File: (select your .sh file from dropdown)

From here you can adjust the Thresholds to Alert on the Status of the monitor (which is a 0-100 percentage of the actual Free Memory available).

Enjoy!

See Also: Custom Memory Monitor Part 2

Debugging .hprof files using the Memory Analyzer Tool

If a SiteScope GUI becomes completely blank, or is otherwise unavailable, one of the culprits may be a lack of available disk space – on account of SiteScope creating a number of large JVM debugging files (.hprof files).  Over 35GB’s worth is not out of the question, and this can clearly tank an application as a result.  Essentially, these files are memory dumps that indicate a problem with the application configuration.

To open and parse these files, use the Eclipse Memory Analyzer Tool (MAT).
Download the zip file and extract its contents into the same SiteScope directory where javaw.exe lives:  C:\SiteScope\java\bin

From here on you are on your own, but the basic gist of it goes like this:

  • Open the MAT .exe file, which is located in the same directory you extracted the zip into.
  • Import/Open the SiteScope .hprof files, which are located in SiteScope bin folder: C:\SiteScope\bin
  • …And run all of the canned analysis reports against the .hprof files as you wish

Unless you are a Java Developer, the memory analyzer will not be terribly useful to you.
However, MAT will automatically zip up every Report you generate – which is a perfect format to forward to HP Support, along with the raw .hprof files themselves.  The zipped Reports should end up in the SiteScope\Bin directory after being created.

As a final note, the MAT *should* work against any other HP BTO product that produces .hprof files, such as BSM, however I have not tested this and would only do so after performing a full file system/db backup.

Enjoy!

Return top