| Solstice Enterprise Manager 4.1 Management Information Server (MIS) Guide |
|
Operating MIS
This chapter provides instructions for starting, monitoring and shutting down the Solstice Enterprise Manager (Solstice EM) MIS.
This chapter describes the following topics:
2.1 Starting MIS
MIS starts automatically when you invoke the em_services command. Upon startup, the database server is initialized before any of the MIS daemons.
Provided a license file is installed correctly and a license is available, by default, MIS starts automatically when your machine is rebooted. To start MIS on your own, remove the following files:
- /etc/rc2.d/S96mis
- /etc/rc2.d/K96mis
When you invoke the em_services -start command the appropriate Solstice EM server daemons are started. (Refer to the Managing Your Network for a complete description of Solstice EM daemons.) None of the applications start when you invoke the em_services -start command.
2.1.1 After a New Installation
If you chose not to have the em_setup program automatically start MIS, you must use the em_services -reload command (reinitializes the platform) to start MIS after the installation. Use the procedures in Section 2.1.6 Recreating the Database."
Installation can also be done by using the Graphical User Interface (GUI) called Setup.
2.1.2 Under Normal Conditions
Use the following procedure for starting or restarting MIS under normal conditions.
To Start or Restart MIS
1. Login as root.2. Type:
- /opt/SUNWconn/bin/em_services -start
Note – If you installed the product in a different directory, substitute your partition name for /opt.
2.1.3 After a System Failure
If your system has a failure, restart Solstice EM using the em_services -start command. The database server automatically detects the failure and attempts to recover. During recovery, the database rolls back any uncommitted transactions.
If the database cannot recover, or your system does not restart, you must recover the database from backups. For instructions, see Chapter 5.
If the problem with your system and database is not resolved, you may need to repair or recreate the database. For instructions, refer to either Reintializing the Database" or Recreating the Database."
2.1.4 After a System Reboot
If the system has crashed and been rebooted with the em_services -start command, MIS is designed to start without any data loss.
1. If MIS fails to start, type:
- /opt/SUNWconn/bin/em_services -reload
Caution – The following procedure permanently removes all data from MIS, including managed objects and discovery data. Always make backups in case you have damage later on. Also, it is strongly recommended that you export topology information to a file. Additionally, you may want to export request templates, conditions, poll rates, and severities. You may, also export log data for later input.
Note – If you installed the product in a different directory, substitute your partition name for /opt
2. Initialize the system by typing:
- /opt/SUNWconn/bin/em_services -init
2.1.5 Reintializing the Database
Use the following procedure to repair a damaged or corrupted database. Also, you may want to use the following procedure when you have made numerous changes to the configuration, then decide you want to start over with a clean configuration and database.
Caution – The following procedure deletes data from database tables and MPA tables. If you have not made a backup of the data, you should do so before deleting it. Also, it is strongly recommended that you export your topology data to a file for later importing into the database. For instructions, see Chapter 5. Additionally, you may want to export request templates, conditions, poll rates, and severities. You may also want export log data to input later. For more information on using the
import/export tool, see Chapter 5.
To Reintialize the Database
1. To delete data in database tables, type:
- /opt/SUNWconn/bin/em_services -init
Note – If you installed the product in a different directory, substitute your partition name for /opt.
- This command cleans the database and restarts Solstice EM.
2. When prompted to delete all current information, type:
- `Y' to continue, or `N' to return to the shell prompt.
- When you type Y, the system deletes runtime data. If you type N, the system cancels the command and returns to the shell prompt.
3. When prompted to delete all historical information, type:
- `Y' to continue, or `N' to return to the shell prompt.
- When you type Y, the system deletes all historical log tables. If you type N, only the current information is removed. The historical information remains.
2.1.6 Recreating the Database
Use the following procedure to remove all data in your database. This procedure completely recreates your database and recompiles the Management Information Base (MIB) files.
For example, you may want to use this procedure after you load new object models and want to start with a new database and configuration, such as would be available by the more time consuming process of reinstalling Solstice EM.
Caution – The following procedure permanently removes all data from MIS, including managed objects and discovery data. If you have not made a backup of data, it is recommended that you do so before deleting any data. Also, it is strongly recommended that you export topology information to a file. Additionally, you may want to export request templates, conditions, poll rates, and severities. You may also want export log data to input later. For more information on using the
import/export tool, see Chapter 5.
To Remove all Data in Your Database
1. Start MIS, recreate the entire database, and recompile all MIBs, GDMO, and Abstract Syntax Notation One (ASN1) documents by typing:
- /opt/SUNWconn/bin/em_services -reload
Note – If you installed the product in a different directory, substitute your partition name for /opt.
- A prompt asks if you want to delete all data from the database.
2. When prompted to delete all database information, type:
- `Y' to continue, or `N' to return to the shell prompt.
- If you type Y, the system performs the following actions:
- stops the database server and removes all data
- recreates the directory and database structure
- recompiles and reloads MIBs, GDMO, and ASN1 documents
- recreates database tables and indexes
- restarts Solstice EM daemons
2.1.7 Options and Commands for em_services
This section summarizes all the options and commands available for the em_services command. The following table lists the options available for the em_services command. You can use multiple options as required.
To Invoke the Help Option
- Invoke the help option by typing:
- /opt/SUNWconn/bin/em_services <command>
Note – In order to run commands from the command line, you will first need to source the emenv.csh (emenv.sh) file.
The following table lists the commands available for the em_services command. Unlike the options for the em_services command, only one command can be specified at a time.
The following table lists the options available for the em_services command. You can use multiple options as required.
TABLE 2-2 Optional Parameters for the em_services Command -debug Enables debugging -force Forces the command to work without prompting you for confirmation -quiet Quiet suppresses all output
2.2 Monitoring the Daemon Processes
The Process Monitor monitors the daemon processes started by the Solstice EM platform. If a process dies unexpectedly, the monitor will attempt to restart the process. Some major daemons such as em_mis cannot be simply restarted, since they are critical to the functioning of the platform. In these cases, the platform is restarted by the monitor(em_services -start).
Depending on their design, applications may be able to deal with the timeout period while a process is being restarted. If not, then the applications may need to be restarted.
The process monitor also provides a facility for escalating actions in the case of repeated failure. If a process failure results in the platform failing, then the platform is restarted. If the restart is unsuccessful, then a pre-defined escalation action can be performed. Up to 4 levels of escalation are supported which are user configurable.
Note – You must logon as root to perform the actions of starting and stopping the monitor.
2.2.1 Starting the Process Monitor
The process monitor can be started either of these ways:
- The monitor is automatically started when you invoke em_services command to bring up the Solstice EM platform.
Note – Choose no-monit option to start the process if you do not want to start automatically.
- You can start the process monitor when the Solstice EM platform is up and running, by running em_monitor -restart, which will terminate any running monitor and start a new one. If the Solstice EM platform is not up, then you must bring it up by running em_services with the appropriate option.
Note – You cannot start a monitor manually while em_services startup script is still running while the platform startup is in progress.
2.2.2 Stopping the Process Monitor
The process monitor can be stopped in either of these ways:
- Is automatically stopped when the em-services -stop command is executed.
- You can stop by running em_monitor -stop command, which will stop all the monitoring activity. This command also cleans up and initializes the escalation level back to 0, in case recovery fails and manual intervention is required.
2.2.3 Monitor Timeouts
When em_services start, the monitor monitors the progress of em_services. If any daemon does not start within a specified time period, the monitor assumes that it has failed and will initiate corrective action.
In some cases, such as reloading a platform ( with em_services -reload), with a large number of objects, the monitor may timeout the process, and erroneously initiate recovery actions.
In order to avoid this situation, you can do one of the following:
- Use the no-monit option with em_services while doing a -reload or -init. Once the Solstice EM platform is up, you can start the monitor by running em_monitor -restart.
- You can increase the timeout period by setting the environment variable MONITOR_DELAY. The default value of this variable is 20, for example, by setiing MONITOR_DELAY variable to 40, you can double the timeout period.
Note – MONITOR_DELAY can only be set to an integer value between 1 and 1000. You do not need to change MONITOR_DELAY to account for system speed.
2.2.4 Configuring Escalation Actions
You can specify any action to be performed by the monitor when the platform fails. There are four levels of escalation, 1 to 4. Level 0 is only a process restart and not user configurable. Level 1 is the first user-configurable level, and performed when level 0 fails. If level 1 fails, the monitor escalates to level 2, and so on, until level 4 is reached. If level 4 is reached and the platform is still not up and running, then the level 4 action is repeated endlessly. If the monitor is stopped at any time, then all monitoring and escalation activity ceases.
The escalation action to be specified can be any executable Unix command or script. You specify the action in a file in the /opt/SUNWconn/em/bin directory called escalation_info. The format for specifying the action is:
<level> <action> <level> <action><level> is a number from 1 to 4, and <action> is any Unix command or executable shell script file name. <action> must be specified on a single line, ("\" is not allowed). Each <level> <action> pair should be on a separate line. Any whitespace before <level> is ignored. If level is not a number from 1 through 4, the entire line is ignored. A line beginning with a "#" is a comment line and is ignored.
There are system specified defaults for each escalation level. These are:
# System defined escalation levels
1 em_services -start 2 em_services -stop -force 3 em_services -stop -force 4 em_services -stop -force
Note – em_services -stop (or -abort) will stop the monitor as well.
Consider this example, Supposing you want to stop the platform at level 3 and send a message to Fred, the administrator. The file escalation_info would be written as:
Where call_fred is a shell script:
CODE EXAMPLE 2-1 File escalation_info
# Level 3 escalation: Need to wake up Fred. 3 call_fred
#! /bin/ksh /opt/SUNWconn/em/bin/em_services -stop -force mail fred@home < panic_message When level 3 is reached, this would stop the platform (and all further monitoring activity), and would execute the mail command. All the other levels would be specified at their default levels.
Note – Any escalation action/script should either stop the platform or attempt to restore the platform. Otherwise, the monitor will escalate to the next level.
Caution – When specifying em_services -init or em_services -reload as actions, remember that these actions could destroy ALL runtime data and should be used with care.
2.3 Shutting Down MIS
The two methods for shutting down MIS are normal mode and emergency mode. The first method is the standard for most cases, providing a graceful shut down process. The second method is for emergencies, when you need to shut down MIS immediately to prevent data loss or corruption.
2.3.1 In Normal Mode
Use the normal mode to gracefully shut down MIS and any processes. The database implicitly rolls back all active transactions and immediately disconnects all users. Pending processes are placed in shutdown mode, and when conditions are acceptable, the processes are exited.
Caution – Do not use the kill command to shut down any Solstice EM server processes.
To Shut Down MIS in Normal Mode
- Shut down MIS in normal mode by typing:
- /opt/SUNWconn/bin/em_services -stop
Note – If you installed the product in a different directory, substitute your partition name for /opt
2.3.2 Using Emergency Mode to Shut Down MIS
Use the emergency mode to immediately shut down MIS and any processes. Always try to perform a shutdown using the normal mode method before advancing to the emergency mode method.
Use the emergency mode for shutting down MIS when:
- MIS is functioning irregularly and you have already tried the normal shut down method.
- You experience problems while starting MIS.
- You need to terminate MIS immediately due to an emergency, such as an impending power outage.
To Shut Down MIS in Emergency Mode
- Shut down MIS in emergency mode by typing:
- /opt/SUNWconn/bin/em_services -abort
Note – If you installed the product in a different directory, substitute your partition name for /opt
- Once you enter the abort command, MIS stops immediately.
|
Sun Microsystems, Inc. Copyright information. All rights reserved. |
Doc Set | Contents | Previous | Next | Index |