Halcyon Monitoring Solutions Home Page 
 
Products  |  Services  |  Downloads  |  News  |  Support  |  Careers  |  About Us  |  Contact Us  

Support
Technical Support
Documentation A-Z
FAQ
Halcyon Forums
Ask an Expert
Submit an RFE
Daylight Savings Time Update




FAQ
All Questions
Halcyon Products FAQ
Sun MC 3.x FAQ
Sun MC 3.0 FAQ
Sun MC 2.x FAQ




Learn More About
Service Management Desktop-To-Grid™

Gear into Sun Management Center 4.0

Solaris 10 Container Appliance

Managing Solaris 10 on x86

IBM Tivoli Solutions

VMware Solutions

How to Buy

Subscribe to our Newsletter








  Home >> Support >> FAQ >> Sun MC 2.x FAQ




Sun Management Center 2.x FAQ

Product Questions

1.1 What are Sun Management Center's hardware requirements?

The following system requirements are documented in the Sun Management Center datasheet which may be view at http://www.sun.com/solaris/sunmanagementcenter/docs-2.1.1/.

For the Console, the minimum and suggested hardware requirements are:

Minimum - SPARCstation 10 or Windows NT/98/95 workstation with 96MB RAM
Suggested - Ultra 1 or Windows NT/98/95 workstation with equivalent or better CPU power with 256MB RAM

For the Server Layer, the minimum and suggested hardware requirements are:

Minimum - Ultra 2 machine with 128MB of RAM with 128 MB of swap space
Suggested - Ultra 2 or machine with equivalent or better CPU power with 256MB RAM with 200 MB of swap space

To get better performance from both the Console and Server Layer running on systems that meet the minimum requirements, Halcyon recommends running the Console and Server Layer on different hosts.

1.2 What operating systems does Sun Management Center support?

For the latest information on supported platforms, visit http://www.sun.com/solaris/sunmanagementcenter/docs-2.1.1/.

Halcyon has announced the availability of Sun Management Center inter operable PrimeAlert Agent for Windows NT and PrimeAlert Agent for HP-UX. A version with Linux is now also available here support will be released soon.

1.3 What processes are part of Sun Management Center?

On a SunMC server these processes are running:

   Agent process: esd - init agent -dir /var/opt/SUNWsymon

   Config server process: esd - init cfgserver -dir /var/opt/SUNWsymon

   Topology process: esd - init topology -dir /var/opt/SUNWsymon

   Trap process: esd - init trap -dir /var/opt/SUNWsymon

   Event process: esd - init event -dir /var/opt/SUNWsymon

   Lock Manager process: ./lm_ip -z /var/opt/SUNWsymon/db/eventmgr/log/ -u 200

Java Server process:

   /opt/SUNWsymon/jre1.1.6/bin/bin/sparc/green_threads/jre -DINTERFACE_PATH=/va

Agent host has the agent process running:

   Agent process: esd - init agent -dir /var/opt/SUNWsymon

And when a user starts a GUI this process will appear on that host:

   /common/jdk1.1.6/bin/bin/sparc/green_threads/jre com.sun.symon.base.console

Note: Path /common/jdk1.1.6... depends on the location of jdk installation.

1.4 What modules are available for Sun Management Center?

The modules bundled with the Sun Management Center framework are described below:

Agent Statistics
monitors the health statistics of Sun Management Center Agents


Config-Reader
monitors hardware on the host including memory errors, board temperatures, power supply status, etc.


Directory Size Monitoring
monitors the size and growth rate of a directory and its subdirectories


File Monitoring
monitors the size, growth rate, and time of last modification of files


MIB-II Instrumentation
provides MIB-II data as per RFC 1213


Kernel Reader
monitors kernel statistics such as system load, CPU, file system usage, memory usage, swap usage, users logged in, and several other features


NFS Filesystems
monitors the disk space of NFS filesystems


NFS Statistics
monitors RPC and NFS statistics for the host


Simple MIB-II
displays information related to the systems and interfaces groups of MIB-II nodes


Solaris Process Details
monitors process statistics for the Solaris OS


Data Logging Registry
displays information on all data being logged by the agent


Dynamic Reconfiguration
monitors the dynamic reconfiguration properties (only available for specific platforms)


Health Monitor
monitors the health of the host and offers suggestions on corrective action when alarm conditions occur


Print Spooler
monitors the status of the printer daemon, print queue, and print devices on the local host


Process Monitoring
monitors process characteristics: Process Count, % CPU Usage, virtual size, resident set size, etc.


File Scanning
scans files for user-defined regular expressions. This can be any text file or Sun Management Center circular log file either user-defined or system.


HP JetDirect
monitors printers with the HP JetDirect card


MIB-II Proxy Monitoring
monitors MIB-II parameters for remote SNMP Agents

Appendix C in the Sun Management Center documentation contains more specific descriptions of each module.

The Sun Management Center Partner Forum (http://www.sun.com/sunmanagementcenter/partner-forum) lists companies providing software which integrates with Sun Management Center.

Add-on modules that extend the Sun Management Center capabilities are available from Halcyon Monitoring Solutions Inc. To see all of the add-on products currently available click here.

1.5 How much CPU/memory can I expect the SunMC Agent and Server processes to consume?

Resource consumption varies widely with the level of monitoring you are doing. This is a list of "average" resources each component uses out-of-the box. As the SunMC server starts monitoring additional hosts, or as the Agent has more modules loaded, more resources will be consumed.(Resident Set Size, % CPU)

SunMC Server Processes:

   Topology Agent   9.0MB   1.0
   Security Agent   6.5MB   <1.0
   Trap Agent   5.5MB   <1.0
   Event Manager   7.5MB   <1.0
   Lock Manager   0.8MB   <0.1

SunMC/PrimeAlert Agent Process:

   Solaris   8.0MB   1.0
   NT 4.0   6.0MB   1.0
   HP-UX   5.5MB   1.0
   Linux-UX   7.5MB   1.0

Note: Solaris statistics are taken from an Ultra5 (Solaris 7), NT4 statistics from a Celeron 400 (SP5), Linux statistics from Pentium 2 400 (RedHat 6.1 2.2.12-20smp), and HP statistics from a 712/60 (HP-UX 10.20)

General Questions

2.1 The Java console seems to be slow. Any suggestion?

Some performance improvement has been noticed when using the Windows NT console. You could also try using the PrimeAlert WebPortal module which provides read-only access (plus alarm acknowledgement) using a standard web browser.

2.2 How can I monitor the server components and ensure they are always running?

You need to run another server (secondary) on a different host which will monitor the first server (primary) and vice versa. In the secondary host add objects which will monitor the primary server components. Use 'Create Topology Objects' to create primary component objects. Enter the hostname and port number of the components (agent - 161, trap - 162, event - 163, topology - 164, config server - 165, server - 2099). Now for each primary object add an alarm action (right click on the object and select 'Alarm Action...') - alarm notification through paging, email etc. Do the same for primary server. Now in case any of the server components on either the primary or secondary server goes down, you will get alarm notification.

2.3 I moved an Agent to a different server context. How can I remove the remaining context information from the first server?

Sun provides an interactive script to manipulate context information, including removing entries (like when an Agent changes servers). The script is called es-servercontrol.sh and can be found in /opt/SUNWsymon/base/sbin. When you run the script as root you will be given a list of options to choose from. The one you want is:

   8) Remove a host:port from the Cfgserver Engines Table

This will remove the remaining cfgserver information from your initial server.

2.4 Does Sun Management Center support any of Sun's storage arrays? If so, how do I get array information from within my console?

Sun Management Center currently only supports the A5x00 series of arrays. Disk information will show up under Operating System -> Kernel Reader -> Filesystem Usage. The physical view can be accessed through the Hardware tab by selecting the socplus card, right-clicking, and selecting the "sena" option.

This is a bug in the swap rule for the Health Monitoring Module in Sun Enterprise SyMON 2.0. This bug has been fixed in version 2.0.1. You should obtain the latest version of the Sun Management Center framework from Sun Microsystems.

Module Questions

3.1 Why is the Health Monitoring Module always telling me that I have too much swap?

This is a bug in the swap rule for the Health Monitoring Module in Sun Enterprise SyMON 2.0.1.

Troubleshooting Questions

4.1 What can I do to make Sun Management Center run faster?

Sun Management Center will run better if the Console and Server are run on separate hosts. This is especially true for users running Sun Management Center on hosts that do not meet the minimum requirements. For users running Windows NT/9x/2000, the Sun Management Center console can be installed and run locally.

In general, CPU and memory resources consumed by the Agent may be decreased by unloading unnecessary modules from the Agent. This can be done directly from the Console by selecting the module in the left hand side hierarchy pane, clicking right and selecting "Unload Module" from the popup menu.

Alternatively, the active module time window may be used to automatically disable modules when they are not required using the Attribute Editor. See Sun Management Center documentation for more information. Disabled modules do not consume CPU but still consume memory.

Note: Console updates (e.g. alarm status changes) may appear to be slow if Agents are not configured to send SNMP traps to the Server Layer host. In that case, consoles will only update on a polling basis, rather than updating asynchronously when traps are sent, e.g. when alarms occur. The Server Layer host is specified when installing the Agent.

The level of security used will also affect how quickly data appears in the console. Strong security will result in slower than usual console updates. Medium security is recommended for users looking to improve the speed of console updates.

In addition if you are using Sun Enterprise SyMON 2.0.1 you can try this (from the SyMON 2.0.1 Release Notes - bottom of page 9):
"The Attribute Editor for the data properties of the system object (in the System folder of the Config-Reader module in the Details window) has a default refresh interval of 10 seconds. We suggest you change the default value to 300 seconds unless you are using the Dynamic Reconfiguration feature.."

4.2 What do I do if the Sun Management Center console doesn't appear on my Windows NT/9x/2000 desktop?

If you start a Sun Management Center console from Windows and the login window does not appear on your desktop, check the Taskbar at the bottom of you desktop for the Management Center task. Use this button to access the login window by maximizing it and resizing it accordingly.

4.3 What do I do if the console hangs during login?

The console will hang during login if the Server Layer is unable to resolve the host name of the host on which the console is being started. In this case, the Server Layer host's local /etc/hosts file or the map in NIS / NIS+ will have to be updated.

When editing the host database, the Server Layer must be restated for the change(s) to take effect.

In addition if you are using Windows console and the problem hasn't been solved edit the file "C:\Program Files\SyMON2.0\esymon\es-console.bat" on the console host to include its ip address.

Change the line:
"\Program Files\SyMON2.0\javasoft\bin\jre" -c ...
to
"\Program Files\SyMON2.0\javasoft\bin\jre" -Djava.rmi.server.hostname=<ip_address> -c ...
where <ip_address> is the address of the NT console host.

4.4 What do I do when I get the message "Unable to load console info" when trying to open the details window for my Agent?

This is a symptom of a bug in Sun Enterprise SyMON 2.0 and 2.0.1 when the Agent is not running. Check to see if the Agent is running by entering "ps -ef | grep agent". There should be a process named "esd - init agent" If the Agent is not running, restart it:

  1. Login as root on the Agent Host
  1. Enter "cd /opt/SUNWsymon/sbin"
  1. Enter "./es-start -a"

4.5 When I bring up the module list window for the Agent on the server layer, why do I see a module that was installed on another system? When I try to load this module, I get "Module load failed" in the 'Module Parameters' window?

This issue is only on Sun Enterprise SyMON 2.0. To support SyMON 2.0 console/server layers, the module parameter file is installed on the SyMON 2.0 server layer. The Agent on the server layer is able to read the parameter file and is tricked into thinking that it can load the module. However, since the remaining module definition files were not installed on this system, trying to load the module will fail.

If you wish to be able to load this module on the Agent running on the server layer, install the appropriate Agent package on this system. See the README.install of the module you wish to install for more information.

4.6 What do I do when I try to create an object and get the message "Create denied. Check online help for license info."?

The Sun Management Center framework available from www.sun.com/sunmanagementcenter is licensed to monitor one host. To monitor more hosts, additional licenses must be purchased. Information on purchasing licenses is available at www.sun.com/solaris/sunmanagementcenter/get.html.

Purchased licenses are installed using the script 'es-lic' located in '/opt/SUNWsymon/sbin'. Once installed, make sure you restart the topology server using './es-stop -p' and then './es-start -p'.

This problem occurs only in version Sun Enterprise SyMON 2.0.

4.7 What do I do when I get the message "Cannot find agent process. Please check host/port." when I try to create a node for my Agent?

First, verify that the Agent process is running. Log in to the Agent Host and enter:

'ps -ef | grep agent'

The following should be returned:

esd - init agent -dir /var/opt/SUNWsymon

If the Agent is running, verify that the correct port is specified:

  1. Log in to the Agent Host
  1. View the file /var/opt/SUNWsymon/cfg/domain-config.x
  1. Check the value of "snmpPort"

If the agent is not running, restart the Agent:

  1. Log in as root
  1. Enter 'cd /opt/SUNWsymon/sbin'
  1. Start the Agent: './es-start -a'

If the Agent process is still not running, the Agent may be aborting execution. Enter 'ps -ef | grep agent' to confirm that the Agent is not running. If it is not, view the Agent log file to ascertain the reason. Sun Management Center log files are located in /var/opt/SUNWsymon/log. Since Sun Management Center log files are circular, 'ccat' or 'ctail' can be used to view their contents. See Question 3.10 for more information on viewing log files.

The last message in 'agent.log' should be "*** aborting execution ***". The previous line to this, should provide the reason for the abort. A common reason for aborting execution is that the port the Agent was attempting to use is already in use. To resolve this problem, kill the process using this port or change the port the Agent uses. Typically, the offending process is the Solstice Enterprise Agent (the process name is 'snmpdx'). See the Sun Management Center documentation for information on resolving SNMP port conflicts.

4.8 I create a topology object, or discover a Sun Management Center Agent, and a 'blue folder' icon appears (same icon when creating a 'General Group' object). Not only is the icon different, but I only see the Info and Browser tabs in the Host Details console.

When a blue folder appears, it is because there is a non-existent family type for this host. This is usually because the agent is returning an invalid type for its Entity Family value. To fix this problem, do the following:

1. Double-click on the blue folder to launch the Host Details window for it, and select the Info tab.

2. Note the Entity Family value from this tab.

3. On the SunMC server:

# cd /opt/SUNWsymon/classes/base/console/cfg

4. Copy a family that closely represents the host machine for which the blue folder appears. For example, if the machine is an Ultra 4 and the Entity Family value from step 2 is sun4u-abc, then do the following:

# cp sun4u-Sun-Ultra-4-family-j.x sun4u-abc-family-j.x

5. Restart the console.

6. Recreate the monitored object.

4.9 Agent suddenly aborts with 'Excessive Virtual Memory Use' in the agent.log file. What is causing this?

This is a known Sun bug that says that MIB-II Instrumentation module may cause the Agent to grow. A workaround is to unload the MIB-II Instrumentation module and load the Simple MIB-II module. If the Agent is still growing unload the Simple MIB-II Module.

4.10 My Event Manager is terminating soon after starting up, and its logfile contains this message: "unable to open lockmgr session".

What is probably happening is that the SyMON lockmanager process was only able to partially initialize itself. This happens when it cannot write a status message to /dev/console. You can verify this is the problem you're having by seeing if a) the lockmanager is running (ps -ef | grep lm_), and b) this file doesn't exist: "/var/opt/SUNWsymon/db/eventmgr/log/evLckMgr".

If you've verified a) and b) then for some reason /dev/console cannot be written to. You can check this by running "echo Testing >> /dev/console" and noticing that this command blocks forever. One possible cause of this is because the permissions on /dev/console have been changed. It is normally "crw--w----". Another reason may be that another process is locking the console for its use, or is writing to the console so fast that no other process can. If you have access to the console you can read it looking for error messages, or a clue to what process may have the device locked. You can also look for a spinning process that may be trying to write to /dev/console.


Copyright © 1996-2008, Halcyon Monitoring Solutions Inc. All rights reserved.
Terms of Use  |  Privacy Policy