Halcyon PrimeAlert Products FAQ

1.0 PrimeAlert for Oracle

1.1 Why does the PrimeAlert for Oracle module have a Down (black) alarm condition with the message "Module Status is Not Accessible" even though the Oracle server is running?

There may be several reasons for this. First, one of the module parameters for the Oracle server may be wrong. Check that the module parameters are correct by running the script HALDBMSAlertOracle-check.sh (included in the distribution), then edit the module and enter the parameters from the script output

Second, you may have encountered a known problem with version 1.1 of HALDBMSAlertOracle-check.sh which was released with PrimeAlert for Oracle version 1.0.1. Newer releases of the module contain the updated script. Please download a newer version of the module.

Finally, you may be encountering a character set problem. Use the following command to find your character set:

SQL> select VALUE from V$NLS_PARAMETERS where PARAMETER='NLS_CHARACTERSET';

Oracle modules version 1.0.5 and earlier do not support multi-byte character sets such as UTF8. This issue is dealt with in later releases. Please download a newer version, or contact Halcyon if you encounter this problem.

1.2 Why does the Agent have Oracle connections open even after the PrimeAlert for Oracle module has been unloaded?

This is caused when one or more processes named pad-oracle do not exit when the PrimeAlert for Oracle module is unloaded or scheduled off. This can interfere with a database shutdown. For example, the pad-oracle process may still have an open connection to the Oracle database. To work around this problem, kill any pad-oracle processes which are still running after the Oracle module has been unloaded or scheduled off.

1.3 What are the minimum privileges required for an Oracle user to load the Oracle module?

The user must have permission to "CREATE SESSION" (which lets them connect to the database), and permission to "SELECT_CATALOG_ROLE". SELECT_CATALOG_ROLE is a pre-defined role in Oracle which enables the grantee to select tables and views in the database dictionary.

1.4 Does your Oracle module support non-ascii (i.e. UTF8) character sets??

The PrimeAlert for Oracle module will monitor databases created with a non-ascii (i.e. UTF8) character sets. However, the named database objects (i.e. tables, columns, segments etc...) and user names that are named with multi-byte characters will not be displayed or monitored correctly.

2.0 PrimeAlert for Sybase

2.1 Can the PrimeAlert for Sybase module be installed on a remote machine instead of on the Sybase server host?

Yes, however the process monitoring and log file scanning capabilities of the module will not work on the remote machine.

To monitor the Sybase server host from a remote machine:

1. Copy the "interfaces" file from the remote machine to a directory on the local machine if the $SYBASE directory on the Sybase server host is not accessible locally.

2. Edit the "Server Directory ($SYBASE)" Database Parameter in the Module Loader to specify the local directory.

3. Edit the DSQUERY parameter used in the "interfaces" file for the remote machine.

2.2 Why is the output for Database Sizes and Transaction Logs sizes different from Sybase stored procedures like sp_helpdb and sp_spaceused?

In these two tables, the space for data and logs is separated as much as possible to allow a clearer picture of your database.

For the Sybase stored procedure sp_spaceused, the 'database_size' field is the sum of all devices allocated for the database, and the 'data' field is the sum of the sizes of all data objects in the database, including transaction logs.

For the stored procedure sp_helpdb, the 'size' field is the size of each device allocated for the database, and the 'free kbytes' field is the amount of space left unreserved on that device.

In the PrimeAlert for Sybase module, the 'Total' field in the Database Sizes table is the sum of all devices specifically allocated for data (devices that sp_helpdb would report as being 'data only' and 'data and log'). The 'Total' field in the Transaction Logs table is the sum of all devices allocated for logs (devices that sp_helpdb would report as being 'log only' or 'data and log'. The procedure sp_spaceused reports the sum of sizes of all devices, whether 'data only', 'data and log', or 'log only'. The 'Data' field in the Database Sizes table is the sum of all data objects in the database except the transaction logs (i.e. syslogs table), while sp_spaceused included syslogs in its calculation of 'Data'. This is also different from sp_helpdb in that sp_helpdb is showing unreserved space on a device, not unused space. The 'Data' field in the Transaction Logs table is the size of the syslogs table. This should be the same as the value reported for 'sp_spaceused syslogs' but differs from the values reported in sp_helpdb because sp_helpdb reports unreserved space, not unused space.

The module's variations from Sybase stored procedures are designed to give the user more precise information by differentiating between data resources (excluding logs) and log resources.

3.0 PrimeAlert for Sybase Replication Server

3.1 Why does the RepServer module have a Down (black) alarm condition even though the Sybase Rep Server is running?

This may be caused when the module cannot connect to the Sybase Rep Server. When the RSSD transaction log is full, the Sybase Rep Server does not accept logins or isql connections. As a result, the RepServer module shuts itself down. You can avoid this situation by monitoring the RSSD transaction log using both the RepServer module and the Sybase module. Thus, if your RSSD transaction log gets close to being full, you can be notified by the modules to take corrective action before your RepServer module shuts down. For example, use the RepServer to notify you when the transaction log is 45, 50, and 60% full as Info, Alert, and Critical alarms; and use Sybase to notify you when the transaction log is 70% full as a Critical alarm.

3.2 I followed the suggestion in Question 3.6 for monitoring the RSSD transaction log, but I still don't receive an alarm notification from the RepServer module before that module shuts itself down. What should I do?

This may be caused by the alarm thresholds being set too high. Suppose you set the alarm to trigger when the RSSD transaction log is 90% full. At one moment, the size of the RSSD transaction log is below the alarm threshold; before the next SunMC refresh interval, the transaction log fills up so that the RepServer module shuts down before it can send the alarm. The solution is to set a lower alarm threshold, such as 70% instead of 90% full, so that the module has time to detect an alarm condition before the log fills up. In addition, you may shorten the refresh interval by displaying the pop-up menu for the RSSD Transaction Log %Used data item, and clicking Attribute Editor. In the Refresh tab of the Attribute Editor, you can adjust the refresh interval.

4.0 PrimeAlert ScriptPack

4.1 Are there any scripts available to check if the host is still ping-able, ftp-able and telnet-able?

PrimeAlert ScriptPack, available for download from Halcyon's website provides a number of useful scripts

5.0 PrimeAlert WebPortal

5.1 What is the PrimeAlert WebPortal?

PrimeAlert WebPortal provides easy, fast and secure access to data, alarms and all the key functionality of the Sun Management Center console using a standard web browser. It also includes additional features such as Advanced Alarm Search and Global Management which give the user the ability to quickly filter alarms, load/unload modules, and set/modify alarm thresholds across entire domains.

5.2 Does the WebPortal support SSL?

Yes, PrimeAlert WebPortal does support SSL. These packages can be obtained from Halcyon Monitoring Solutions by contacting info@HalcyonInc.com

Note: SSL stands for Secure Socket Layer. It is a protocol for transmitting confidential information (such as login passwords) on the Web, SSL uses a private key to encrypt data that's transferred over the SSL connection. Both Netscape Navigator and Internet Explorer support SSL. Please contact info@HalcyonInc.com for information on using SSL with Reporter and WebPortal.

5.3 Is the functionality of the WebPortal the same as the SunMC console?

PrimeAlert WebPortal provides all the main functionality of the Sun Management Center console that is available in the Host Details window. This includes the ability to load/unload modules, add/delete table rows, set/modify security and alarm thresholds.

5.4 What is the difference between the WebPortal and the SunMC Web Interface?

While both use a standard web browser to Sun Management Center, each one does so with a different look and feel. The SunMC Web Interface uses a common look and feel that is similiar to the console. The WebPortal displays status messages and provides direct links to the source of the problem. It also provides Global Management which is similiar to the Group Operations functionality available in the regular SunMC console.

5.5 Can I use my own web server with WebPortal?

No, PrimeAlert WebPortal is packaged and built with the apache web server and will not work without it.

6.0 PrimeAlert Reporter

6.1 What is the PrimeAlert Reporter?

PrimeAlert Reporter maintains historical data of monitored objects and summarizes it in web-based graphical reports. The data can be used to perform long term capacity planning analysis and to measure performance of a system or applications running on that system. In addition, the data can be saved in delimited formats and exported to third party products such as Microsoft Excel. It also allows you to periodically save a list of running processes with their resource statistics.

6.2 What do I need to run the Reporter?

PrimeAlert Reporter will only work with the PrimeAlert WebPortal. The Reporter uses the WebPortal web interface to collect and display monitored data.

6.3 How often does the Reporter poll and store data?

PrimeAlert Reporter polls for data every 2 mins. This data is then collected and stored in a separate bins for long term storage.

The following data bins exist:
Last 24 Hours - Data is collected every 2 mins
Last Week - Data is collected^* from the 24 Hours bin every 20 mins
Last Month - Data is collected from the Last Week bin every hour
Last 4 Months - Data is collected from the Last Month bin every 4 hours
Last Year - Data is collected from the Last 4 Months bin every 12 hours
Last 5 Years - Data is collected from the Last Year bin every day

^*The data is collected using one of the specified algorithms below:
Avg - Average of the data bin
Max - Maximun value in the data bin
Min - Minimun value in the data bin
Last - Last value added to the data bin

6.4 How much space is required to store historical data?

The Reporter requires 350 KB of disk space for each monitored object.

The filesystem that contains the object databases should always have at least 50 MB of free space, or the databases should be moved to a location where more space is available.

6.5 How often does the Reporter process listing poll for data

PrimeAlert Reporter stores a copy of all the processes running on a monitored host every 5 minutes.

6.6 How much disk space is required to store process listings?

The amount of space used for storing process listings depends on the number of processes running on the monitored host. The more processes you have the more memory required to store the listings.

To track each process, about 0.05 MB of disk space is required; thus, for a typical system running about 100 processes, 5 MB of disk space would be required to save the process list snapshots.

6.7 Where is all the collected data from PrimeAlert Reporter stored?

The location of the database files is specified during installation of the PrimeAlert Reporter agent component. The installation default directories are:

   /var/opt/SUNWsymon/db/rrd (historical data for monitored objects)

and

   /var/opt/SUNWsymon/db/ps (process list snapshots)

6.8 Can I use the Reporter for capacity planning?

Yes, the advantages of using PrimeAlert Reporter is its ability to collect and store long term data, and view that data in the form of a graph at any time. This allows you to view trends and the behaviour of your systems and/or applications to forecast future problems. It is also possible to create to correalate various graphed data in a single view.

6.9 Can I use the Reporter to measure system and application performance?

Yes, since PrimeAlert Reporter can collect data for any monitored property, data can be graphed accordingly. This allows you to view trends and the behaviour of your systems and/or applications to forecast future problems.

6.10 Why does the Reporter module show an empty table in the "Objects" section?

When you first load the Reporter module, the "Objects" table is empty. To add monitored objects to this table, click the grey triangle in the row header, and click Add/Delete Rows. A pre-configured list of objects appear in a pop-up window in which you can select the objects for which you want Reporter to log historical data.

Note that the objects in the list are grouped into categories which correspond to the tabs across the top of the Reporter module. When you add an object, it is added to the table corresponding to the tab of that object's category.

6.11 What processes and sub-processes are run for the Reporter module?

On the agent host, the following processes are run:

• esd - init agent
• reporterMain
• pad-reporter-ps
• pad-init gateway

• /var/opt/SUNWsymon/HALWebPortal/bin/jre.vfb
• Xvfb

6.12 Do I need to upgrade any other modules when I install Reporter?

There are new versions of the Oracle, Sybase, Sybase RepServer, and Cisco modules that are released at the same time as Reporter. Let's say you choose to continue using the older version of the Oracle module; you will see in Reporter that you will not be able to tell the difference from one instance of an Oracle sub-module and another. Similarily for the Sybase, Sybase RepServer, and Cisco modules.

The newest versions of these four modules do not have this problem. If you choose to upgrade these modules, you will have to unload the old modules first, install the new versions, and then load the new modules. Note that when you load the new modules, your alarm thresholds associated with the old modules will not be preserved in the new module, and that you will have to set them again.

6.13 How do I disable the feature of "periodically saving a list of running processes with their resource statistics"?

With the WebPortal process lists feature, any SunMC user can view detailed information on the processes currently running on a monitored machine. This can be of concern if sensitive information is part of the process arguments, ie. passwords.

The process lists can be disabled, to do so:

- Add the following entry to the HALReport-user-config.x file in ESDIR/cfg:

psConfig:enabled = OFF

Note: Access to previously stored process list information will also be refused.

7.0 PrimeAlert EventAction

7.1 What is PrimeAlert EventAction?

PrimeAlert EventAction centralizes the handling of events from all agents in a Sun Management Center configuration and performs user-configured actions for events of interest. It can also be extended to integrate Sun Management Center events with 3rd party products such as HP Openview and Micromuse Netcool.

7.2 Why do I need PrimeAlert EventAction?

PrimeAlert EventAction is essential if you are planning on configuring actions for general events. For example, to setup email notification for every "Warning" and "Critical" alarm in Sun Management Center without EventAction would require the user to drill down into each agent, locate the property and specify the particular action. This could potentially take a long time with many hosts and modules loaded. With EventAction the action only has to be specified at the server level. Events can be filtered so that actions are taken on particular events.

7.3 Where do I install PrimeAlert EventAction?

PrimeAlert EventAction only needs to be installed on the Sun Management Center server. Since all the event processing is done on the server the actions have to be place on the SunMC server machine in following directory:
/var/opt/SUNWsymon/bin/HALEventAction/

7.4 How do I setup alphanumeric paging in Sun MC?

The simplest way to setup paging is to create an email account that sends out pages. This is a typical feature for recent alphanumeric pagers, and can be activated by contacting your paging company. Paging can also be setup through a modem, Halcyon Monitoring Solutions provides scripts to do so, contact info@HalcyonInc.com.

7.5 How do I setup EventAction to avoid receiving repeated alarms?

Sun Management Center triggers an event everytime a rule is satisfied. For example, if a monitored process consumes 90% of the cpu this triggers an alert alarm. When the process cpu falls below 90%, the event is closed and a new one is issued everytime the process cpu is greater than 90%. This sort of behaviour triggers an action more often than one is needed. EventAction allows you to set a deadband or a delayed alarm before an action is taken. In our example above, an action would only be taken if the event is still open after some specified time. The same feature can be used to escalate alarms that are not fixed by a specified period of time allowing notification to go from the user to management level.

7.6 Can I use EventAction to escalate problems not fixed within a given time?

7.7 How do I setup EventAction to send email notifications with direct links to events?

PrimeAlert EventAction can provide direct links to events when used with the PrimeAlert WebPortal. When loading the EventAction module, specify the WebPortal URL and make sure to specify HTML based email notification. This will send you event emails with direct links to the particular event.

7.8 How do I forwards Sun MC events to third party products

With PrimeAlert EventAction, Sun Management Center events can be passed to 3rd party SNMP compliant tools. Currently Halcyon has adapters to forward events into PrimeAlert Adapter for IBM Tivoli, PrimeAlert Adapter for HP OpenView, and PrimeAlert Adapter for Netcool.

Halcyon also has a generic utility to pass Sun MC events into other products. For more information, please contact Halcyon For the most up-to-date list of PrimeAlert Adapters currently available, please check out our downloads page.

7.9 I would like to do integration of status traps with an external manager. How can this be done

Halcyon has a generic utility to integrate Sun Management Center traps into other management platforms. For more information, please contact Halcyon

7.10 Sometimes my PrimeAlert EventAction module seems to trigger multiple events, even for old or closed alarms. What is going on?

This is a known bug with Sun's Event Manager (one of the server process) in SunMC 2.x that can cause it to re-read old alarms from an Agent. This has been fixed in the lastest 3.0 release of SunMC.

7.11 If an event occurs that matches more than a single action what happens?

EventAction runs all actions that are triggered by an event. There is currently no option to rank actions such that only the first matched action is run.

7.12 Does EventAction run action on the SunMC Server, or the SunMC Agents?

EventAction runs actions on the SunMC Server only. It does not communicate back to the Agent where the event occured to run the action you choose.

7.13 I do not own PrimeAlert FileBrowser. Can I still view the event history?

Yes. On your SunMC Server run this command:

/opt/SUNWsymon/sbin/es-run ccat
/var/opt/SUNWsymon/log/EventActionHistory.log

This will dump the event history to standard out. Redirect to a file as needed.

7.14 Why doesn't EventAction work after I apply the lastest Solaris patch for SunMC?

When a SunMC 3.0 server host is used, PrimeAlert EventAction may not detect any events. This occurs when the following Solaris patches are applied:

110971-02 (Solaris 2.6)
110972-02 (Solaris 2.7)
110973-02 (Solaris 2.8)
110936-02 (Solaris 2.6)
110937-02 (Solaris 2.7)
110938-02 (Solaris 2.8)

This has been fixed in a patched version of PrimeAlert EventAction. Please download the latest release if you have applied one of the patches listed above.

8.0 PrimeAlert for Cisco Routers

8.1 What Cisco devices are supported by the PrimeAlert for Cisco Router Module

PrimeAlert for Cisco Router supports various devices. Each monitored device depends on the information available in the MIB for that particular device. The easiest way to determine if a device is supported is to load the module with the device of interest and see if the information updates properly.

The following devices are supported:
as5100 as5200 as5300 as5800 c1001 c1002 c1003 c1004 c1005 c1020 c12000 c1601 c1602 c1603 c2000 c2102 c2202 c2500 c2501 c2502 c2503 c2504 c2505 c2506 c2507 c2508 c2509 c2510 c2511 c2512 c2513 c2514 c2515 c2516 c2517 c2518 c2519 c2520 c2521 c2522 c2523 c2524 c2525 c2610 c2611 c3000 c3101 c3102 c3103 c3104 c3202 c3204 c3620 c3640 c4000 c4500 c4700 c6200 c6400nsp c7000 c7010 c7200 c7505 c7507 c7513 c8510csr cs500 igs ls1010 mc3810 ubr7200 wsc2900xl wsc8510csr wsc8510msr wsc8540csr wsc8540msr

9.0 PrimeAlert Agent for Linux

9.1 I don't have RedHat Linux 6.0 or 6.1. Will it still run

Probably. Halcyon uses RPM as the installation utility, so other distributions (i.e. Mandrake) that use RPM should work as well. There are also utilities that convert RPM files to .tar archives, but installation by this method is unsupported.

9.2 Linux runs on Alpha/PowerPC etc.. Are these architectures supported?

No. Currently only x86 architectures are supported. Contact Halcyon about ports to other platforms.

9.3 SunMC shows hardware information for many Sparc hosts. Does your Linux Agent show hardware information as well?

No. The wide variety of x86 compatible hardware on the market makes this infeasible. It is possible for Halcyon to do integration with other pacakges such as lmsensors. Contact us directly for more information.

9.4 Is the Linux Agent GPL?

Although Halcyon sells products containing GPL code, and complies with the license accordingly, the Linux Agent itself is under a proprietary license.

10.0 PrimeAlert EventAction AlarmRouter Pack

10.1 Can the AlarmRouter pack be used to forward SNMP traps to frameworks other than HP Openview or Micromuse Netcool?

The AlarmRouter package will install a /var/opt/SUNWsymon/cfg/HALEventTrap.mib file on your SunMC Server. This files describes the format of the SNMP traps that EventAction sends out. This MIB file can be used with other frameworks so they understand SunMC events.

10.2 Does the Halcyon PrimeAlert AlarmRouter Pack allow me to integrate events into BMC Patrol or NetIQ?

The AlarmRouter pack contains a generic SNMP trap forwarding mechanism that can be used to forward Sun MC events to any 3rd party product that can accept traps. We have observed this to work with NetIQ, HP Openview, IBM Tivoli, and Micromuse Netcool. Contact Halcyon for more details.

11.0 PrimeAlert for Microsoft SQL Server

11.1 Why does the PrimeAlert for SQL Server module have a Down (black) alarm even though SQL Server is running?

The module parameters should be verified. The most common problem is the user/password are incorrect or do not have the proper permissions to read the databases (see next question). Also verify the ODBC connection is working properly.

11.2 What user should be used and are there any security concerns? What are the minimum privileges for the SQL Server user to load the SQL Server module?

It is recommended that a system DSN that uses Integrated Security be used, and the Halcyon PrimeAlert Agent service be run under a user that has administrative authority on the MS SQL Server. When integrated security is used, a dummy user name and password should be specified when loading the module.

If the system DSN is not using Integrated Security or the PrimeAlert agent service is not running under a user who has administrative privileges, and you wish to be able to see the input buffer of each running process in the User Activity table, you will have to specify the SA user and password when loading the module. You cannot use a username which is merely aliased to the SA account in the master database.

If you decide not to utilize the above options, and wish to use another internal MS SQL Server user with un-integrated security on the DSN, the following permissions must be granted to the internal database user that you do use:

- In the master database execute privilege on sp_server_info, sp_helpdb, sp_spaceused, and select privilege on the spt_values table.
- The user must be a valid user in the msdb database. When monitoring MS SQL Server 6.5, the user must have execute permission on sp_sysbackuphistory. When monitoring MS SQL Server 7.0, the user must have select permission on the table backupset.
- The user must be a valid user of any other databases if you wish to monitor them, however no specific permissions need be granted.

11.3 The module loaded ok but the Resources table has a black star for the status icon and some of the properties is empty.

The PrimeAlert agent must be run as a user with permission to read the performance registry.

11.4 The total number of user connections is higher than the maximum number of connections. What's wrong?

The total number of user connections includes 4 system processes which don't count towards the maximum.

11.5 Can I monitor remote SQL Servers? (ie. loading the module on a Windows system with the DSN pointing to an MS SQL server on another host)

Yes, but certain data items will not be available. The 'Server Processes' subsection monitoring the related Windows server processes and the 'Error Log' subsection monitoring the MS SQL server logfile will be empty.

Note the PrimeAlert agent must be run as a user which can read the performance registry on the remote host or some tables may be empty and have a black star for the status icon.

12.0 PrimeAlert Agent for Microsoft Windows

12.1 How do I configure the PrimeAlert Agent for Windows using multiple network interfaces?

When a Windows agent is started on a machine with mulitple NICs (Network Interface Cards) it will attempt to use one of the two IP addresses, either the physical or virtual one. Verify the IP addresses using the "ipconfig" command at the DOS Prompt. If your physical card has an IP address and the virtual card does not, the agent will fail to start if the proper binding order is not specified. To change the binding order, follow the steps listed below:

1. Right click on "My Network Places" and select "Properties"

2. Select "Advanced Settings" from the "Advanced" menubar

3. Change the order of interfaces in the "Adapters and Bindings" tab. ie. Promote the virtual NIC over the physical one.

13.0 PrimeAlert for Sun ONE Application Server (formerly iPlanet Application Server)

13.1 The Processes table of the module is generating an alarm on the number of engine processes running (of a specific engine type) even though all the required processes are running. How do I fix this problem?

You have removed an engine from Application Server and the module has not yet noticed this. Because of the load placed on the system, the module will only check for this once per day, unless a manual refresh is forced.

Force a manual refresh of the running Application Server engines. To force a refresh, right click on the main module icon and select "Attribute Editor". Now switch to the "Schedule" tab. Select "Refresh Engines Now" and click OK or Apply.

14.0 PrimeAlert for Sun ONE Directory Server (formerly iPlanet Directory Server)

14.1 Does the PrimeAlert for Sun ONE Directory Server module monitor the Sun ONE Meta Directory product, the SUN One Certificate Server, or the Sun ONE Directory Proxy Servers?

Not at this time. However it should be possible to use a number of other general PrimeAlert and Sun modules for Sun Management Center modules to do some basic monitoring of these products. For example, Sun's Service and Availability Manager's LDAP modules, and Halcyon's PlusPack modules (LogFileMonitor, ProcessMonitor, DirectoryMonitor, etc).

If monitoring of these other Sun ONE products are crucial to your environment, please express your interest using the following Product Survey and then contact us directly to inquire about a possible timeline or custom solution.

14.2 Does the PrimeAlert for Sun ONE Directory Server module monitor replication?

This capability will be released VERY SOON. Please contact us to be notified when it becomes available, and to discuss your requirements.

14.3 What is the recommended method for monitoring a multi-master replication setup?

Load the PrimeAlert for Sun ONE Directory Server module on each host, including all the Masters and all Consumers. In the Main Sun MC topology view, create a Directory Server group. Populate the group with views of each PrimeAlert for Sun ONE Directory Server module from all of the directory server hosts. So your group will contain an icon for each master server and all of the consumers. This provides a top-level view into the health of your entire directory server architecture.

15.0 PrimeAlert for Sun ONE Web Server (formerly iPlanet Web Server)

15.1 What web servers does this module support?

The module currently supports iPlanet Enterprise Web Server 4.1 and Sun ONE Web Server 6.0. Netscape/iPlanet FastTrack Web Server is currently unsupported.

16.0 PrimeAlert for Sun ONE Portal Server (formerly iPlanet Portal Server)

16.1 What modules are packaged with the Sun ONE Portal Server?

The PrimeAlert for iPlanet Portal Server distribution consists of a bundle of the following modules:

- PrimeAlert for Sun ONE Portal Server (formerly iPlanet Portal Server)
- PrimeAlert for Sun ONE PS Gateway (formerly iPlanet iPS Gateway)
- PrimeAlert for Sun ONE Web Server (formerly iPlanet Web Server)
- PrimeAlert for Sun ONE Directory Server (formerly iPlanet Directory Server)

The Sun ONE/iPlanet Portal Server applicaton consists of several components, namely a web server, a LDAP store, a gateway, and the portal server. To monitor these various Portal Server components the user must install and load the appropriate Sun ONE module on each host where these applications are running. These modules have all been packaged together for ease of installation.

16.2 After I load the PrimeAlert for iPlanet Portal Server module, it goes into a "down" state. (Black Alarm). How can I fix this?

The PrimeAlert for iPlanet Portal Server utilizes the iPlanet Gateway "checkport" script usually found in $BASE_DIR/bin where BASE_DIR is /opt/SUNWips by default. This utility is only installed if the iPlanet Gateway component is installed.

Copy the "checkport" utility from the host where the iPlanet Gateway is installed and place it in the same directory on the host where iPlanet Portal Server is installed and running.

17.0 PrimeAlert for Sun ONE Messaging Server (formerly iPlanet Messaging Server)

17.1 What modules are packaged with the Sun ONE Messaging Server?

The PrimeAlert for iPlanet Messaging Server distribution consists of a bundle of the following modules:

- PrimeAlert for Sun ONE Messaging Server (formerly iPlanet Portal Server)
- PrimeAlert for Sun ONE Web Server (formerly iPlanet Web Server)
- PrimeAlert for Sun ONE Directory Server (formerly iPlanet Directory Server)

The Sun ONE/iPlanet Messaging Server applicaton consists of several components, namely a web server, and a LDAP. To monitor these various Messaging Server components the user must install and load the appropriate Sun ONE module on each host where these applications are running. These modules have all been packaged together for ease of installation.

18.0 PrimeAlert for Veritas Volume Manager

18.1 What versions of Veritas Volume Manager are supported in Halcyon's module for Sun Management Center (Sun MC)?

Halcyon has tested the module against versions 3.0 and 3.5, and will support it with any other 3.x version. Please contact Halcyon if you have any other requirements.

18.2 Can I access the raw output of Veritas Volume Manager commands through the Sun Management Center GUI?

Yes, these are provided through standard Sun MC probe commands. Look in the context menu's in the Device table/node, Disk Groups heirarchy node, Disk Status table/node, and the I/O Performance table/nodes. Halcyon's Reporter / WebPortal version 2.0 will contain support for viewing probe commands via a web browser.

18.3 What are the possible values for the Status column in the Devices Table?

"Online", "online failing", "error" and it may contain the string "failed". These states are obtained from Veritas Volume Manager itself, so the alarm conditions and severity that you set should correspond to what actions your team would take should VVM report those states for a disk.

18.4 What are the possible values for the Status column in the Devices Table?

"Online", "online failing", "error" and it may contain the string "failed". These states are obtained from Veritas Volume Manager itself, so the alarm conditions and severity that you set should correspond to what actions your team would take should VVM report those states for a disk.

18.5 What are the Exit Code and Error Messages fields in the "Device Table Update Status" section for?

These will be non-zero and non-null if one or more of the following Veritas Volume Manager commands fails or gives highly abnormal results: VXDISK, VXPRINT, VXSTAT, or VXDG. The specific codes and messages that are possible are defined by the range of return codes and messages which those Veritas commands can return. Thresholds for these two fields are at your discretion.

18.6 Disk Status is listed in two places, in the Device Table in the "Devices" section, and in the Disk Table inside the "Disk Group - Disk Status" section. Where should I configure the alarm?

We would recommend that you place an alarm threshold on the Status field that is in the "Device Table" in the Devices section. Thus if a device is ever moved/shifted from one group to another, the alarm threshold will not be lost. (If the threshold was set in the Disk Group's Disk-Status Disk-Table, and then that disk was later removed from that disk group, that row would disappear from that disk-group, and as a result the threshold would have to be re-set in the new disk-group that the disk had been added to.)

18.7 I am using Veritas Volume Manager in a clustered environment. What is the appropriate or recommended solution for monitoring it in this case?

Please contact Halcyon Monitoring Solutions to discuss our cluster monitoring recommendations.

18.8 How can I see the current alarm thresholds in this module in Sun Management Center (SunMC)?

19.0 PrimeAlert for Veritas NetBackup

19.1 What versions of Veritas NetBackup are supported in Halcyon's module for Sun Management Center (SunMC)?

Halcyon has tested the module against version 3.4 and will support it with any other 3.x version. Please contact Halcyon if you have any other requirements.

19.2 Does the Veritas NetBackup module include any monitored objects that can be used as performance indicators of nightly backups?

Yes, the module displays performance statistics for each backup job, such as the transfer rate (KB/s), as well as the size and time taken to perform each backup.

20.0 PrimeAlert for Veritas Cluster Server

20.1 What versions of Veritas Cluster Server are supported in Halcyon's module for Sun Management Center (SunMC)?

Halcyon has tested the module against versions 1.1 and 2.0 and will support it with any versions inbetween. Please contact Halcyon if you have any other requirements.

20.2 How do I go about setting up the PrimeAlert for Veritas Cluster Server module in a Sun Management Center environment (SunMC)? Do I need to install it on all clustered nodes?

Yes, the module should be installed on each clustered node. This will ensure that network heartbeats between the nodes are active, that each node can access the shared disk and alert users when any kind of failover occurs.

21.0 PrimeAlert for Veritas File System

21.1 Does Halcyon have a Sun Management Center (SunMC) module to monitor the Veritas File System (VxFS)?

Halcyon is currently gathering requirements to support monitoring of the Veritas File System. There is no planned release date. If you would like more information, please feel free to contact Halcyon.

22.0 PrimeAlert Module Scheduler

22.1 Is there a way a specific Sun Management Center (Sun MC) module in a certain host can be turned on and off dynamically by some script/program/command?

Yes, the PrimeAlert Module Scheduler module is designed to do exactly this. It works in a fashion similar to the ScriptRunner module. It executes a specified script and monitors the output of that script. You specify a pattern, when that pattern is matched, the specified module will be disabled. This is useful for turning off modules on holidays so that you do not receive false alarms or events. It also helps to keep the event database cleaner.

22.2 What are some typical applications of this module?

This module is useful if for example a file or process needs to exist during a given time-window. The module or modules responsible for monitoring them can be scheduled for just those time-windows. This module is also useful for configuring monitoring only during business hours, or ensuring that monitoring is turned off on holidays.

Note that the module does not have to follow a "calendar or time based" schedule, as it's state is based upon the output of a script, which can undertake whatever actions you deem necessary to determine whether monitoring should be on or off. For example it can also be used to configure "fail over" monitoring for a backup cluster node. When the primary node goes down or is inaccessible, the modules on the backup node can be configured to turn on.

22.3 Are any scripts provided with the module?

Yes, holiday scripts are provided that will turn off monitoring during either Canadian or American holidays. An IP check script is provided whose status is based on the ifconfig -a output, so this could be used in a cluster environment. Note that these scripts are not officially supported.

22.4 Does ModuleScheduler 1.0.0 supports SunMC 3.0 on Solaris 8 (2.8) or 9 (2.9)?

Yes it does. The old v1.0.0 README is in error. Halcyon will support the above. (Please contact Halcyon for the Solaris 9 version).

23.0 PrimeAlert for Apache Web Server

23.1 What versions of Apache Web Server are supported in Halcyon's module for Sun Management Center (SunMC)?

Halcyon has tested the module against versions 1.3.14 and higher. It does not support Apache 2.x. Apache versions 1.3.19 and 1.3.20 are recommended. Please contact Halcyon if you have any other requirements.

23.2 Does Apache need any special compilation flags to work with the module?

Yes. The Apache Web Server must be configured with Dynamic Shared Object (DSO) support.

i.e. ./configure --enable-module=so

If Apache was not configured with this option it must be reconfigured and recompiled before the module will work.

23.3 Does the Apache Web Server require any "httpd.conf" options to be set to work with the module?

Yes. The module installation script will backup your existing httpd.conf file and write the appropriate PrimeAlert entries to it. Addition/removal of PrimeAlert configuration information can also be done manually using the following commands:

/opt/SUNWsymon/modules/sbin/HALHTTPAlertApache-configureApache.sh ADD
/opt/SUNWsymon/modules/sbin/HALHTTPAlertApache-configureApache.sh REMOVE

23.4 Can this module be used to monitor and report on Apache web server performance?

Yes, the module includes numerous performance metrics on a per-virtualhost basis. Some of the statistics include:

Transaction Summaries:
- page hits in the last 15 minutes, 24 hours, and lifetime
- data served in the last 15 minutes, 24 hours, and lifetime (MB)
- data received in the last 15 minutes, 24 hours, and lifetime (MB)

Document Summaries:
-most popular URLs (15 min, 24 hour, and lifetime totals) based on:
   - page hits
   - total megabytes per URL served
   - total megabytes per URL received

These statistics can all be easily integrated with the PrimeAlert Reporter to generate graphical reports of server load and performance. Reporter also allows the exporting of key Apache web server performance data into .tsv and .csv formats for import into 3rd party utilities.

24.0 PrimeAlert for Network Appliance

24.1 Which Network Appliance systems are supported in Halcyon's module for Sun Management Center (SunMC)?

The PrimeAlert for Network Appliance supports systems running Data ONTAP 6.0 or higher.

24.2. Does this module require any specific settings on the Network Appliance system to function in Sun Management Center?

Yes. The SNMP agent for the device must be enabled. For details on how to enable SNMP for the Network Appliance system, refer to its documentation.

25.0 PrimeAlert for Brocade Switch

25.1 Which Brocade Switches are supported in Halcyon's module for Sun Management Center (SunMC)?

The PrimeAlert for Brocade Switch module currently supports switches running SilkWorm Firmware V2.2 or higher.

25.2. Does this module require any specific settings on the Brocade Switch to function in Sun Management Center?

Yes. The SNMP agent for the device must be enabled. For details on how to enable SNMP for the switch, refer to its documentation.

25.3. Does the module accept SNMP traps from the switch that is being monitored?

Traps are not accepted by the module, however the events table for the switch is scanned. This table represents the traps that have been emitted by the switch.

26.0 PrimeAlert for Hitachi Storage

26.1 What Hitachi Storage systems are supported in Halcyon's module for Sun Management Center (SunMC)?

The PrimeAlert for Hitachi Storage module currently supports 7700E and 9900 systems.

26.2. Does this module require any specific settings on the Hitachi Storage to function in Sun Management Center?

Yes. The SNMP agent for the device must be enabled. For details on how to enable SNMP for the Hitachi Data Storage device, refer to its documentation.

27.0 PrimeAlert InstantPaging

27.1 Is there any way to receive Sun Management Center (Sun MC) event notifications if our email system goes down?

Yes, the PrimeAlert Instant Paging software does not rely on the email system. It transmits numeric/alphanumberic pages via a modem to any TAP (Telelocator Alphanumeric Protocol) device, i.e. Pager.

27.2 If I use the PrimeAlert InstantPaging software, do I still need PrimeAlert EventAction?

Yes, the Instant Paging software is used as an action to particular Sun MC events. The EventAction module is typically used to send out email notifications. Instead, you can specify that you'd like some other means of notification. The EventAction will also allow users to filter events before they are transmitted and keep a central log of events and actions performed.

28.0 PrimeAlert Adapter for IBM Tivoli

28.1 What Halcyon PrimeAlert products are required to integrate Sun Management Center (Sun MC) with IBM Tivoli?

The integration of Tivoli and Sun MC requires the PrimeAlert EventAction and the PrimeAlert Adapter for IBM Tivoli.

28.2 What are the advantages of using the PrimeAlert Adapter for Tivoli from Halcyon as compared to the Sun Management Center (Sun MC) IBM Tivoli integration available from Sun?

The following advantages have been reported by Halcyon customers running the PrimeAlert Adapter for IBM Tivoli:

- The integration is platform independent and does not require the customer to be running IBM Tivoli on a Solaris system.

- The PrimeAlert Adapter for Tivoli keeps state. It does not resend old alarms when the Sun Management Center Server is restarted. As a result, it does not clutter the network operations center with stale alarm messages.

- The PrimeAlert Adapter for Tivoli provides more control to the customer. The user can filter events based on the following criteria:
- Alarm State
- Severity
- Hosts
- Modules
- Status Message
- Time

- The user has the capability to filter out transient events and repeat forwarding of events if they have been left open for too long.

- Most hardware alarms are log events meaning they are opened and closed simultaneously. The PrimeAlert Adapter for Tivoli treats the log events as a separate alarm state rather than a closed alarm and therefore provides more control for the customer. i.e. These events are never lost and can or cannot be forwarded to Tivoli.

29.0 PrimeAlert Adapter for HP Openview

29.1 What Halcyon PrimeAlert products are required to integrate Sun Management Center (Sun MC) with HP Openview?

The integration of Openview and Sun MC requires the PrimeAlert EventAction and the PrimeAlert Adapter for HP Openview.

29.2 What are the advantages of using the PrimeAlert Adapter for HP Openview from Halcyon as compared to the Sun Management Center (Sun MC) HP Openview integration available from HP?

The following advantages have been reported by Halcyon customers running the PrimeAlert Adapter for Openview:

- The integration is platform independent and does not require the customer to be running HP Openview on a Solaris system.

- The PrimeAlert Adapter for Openview handles restarts smoothly. It only sends information for new events. As a result, it does not clutter the network operations center with stale alarm messages.

- The PrimeAlert Adapter for Openview provides more control to the customer. The user can filter events based on the following criteria:
- Alarm State
- Severity
- Hosts
- Modules
- Status Message
- Time

- The user has the capability to filter out transient events and repeat forwarding of events if they have been left open for too long.

- Most hardware alarms are log events meaning they are opened and closed simultaneously. The PrimeAlert Adapter for Openview treats the log events as a separate alarm state rather than a closed alarm and therefore provides more control for the customer. i.e. These events are never lost and can or cannot be forwarded to Openview.

30.0 PrimeAlert Adapter for Micromuse Netcool

30.1 What Halcyon PrimeAlert products are required to integrate Sun Management Center (Sun MC) with Micromuse NetCool?

The integration of Netcool and Sun MC requires the PrimeAlert EventAction and the PrimeAlert Adapter for Micromuse Netcool.

The integration of Patrol and Sun MC requires the PrimeAlert EventAction and the PrimeAlert AlarmRouter Pack. The AlarmRouter pack contains a generic SNMP trap forwarding mechanism that can be used to forward Sun MC events to any 3rd party product that can accept SNMP traps.

31.0 PrimeAlert Neuron Agent

31.1 What is the PrimeAlert Neuron Agent?

The Neuron Agent is a lightweight, fully featured, stand-alone agent that can be easily deployed and configured. Events from Neuron can be directly integrated with standard enterprise management frameworks. The Neuron Agent provides similiar host monitoring functionality (OS and Applications) to the Sun Management Center software without the need for a server component.

The Neuron Agent is also supported on multiple platforms including Sun Solaris, Microsoft Windows, Linux, and HP Tru64.

31.2 How can I forward Neuron events to other 3rd party products (enterprise managers)?

During installation, there is an option to configure the agent to send SNMP traps to a remote host. If this host needs to be changed after the product is installed, or if you would like the agent to send traps to an additional host, a configuration file needs to be modified and the agent restarted. The configuration file is located at:

[INSTALLDIR]\local\cfg\HALEventTarget-config.x (for Windows)
[INSTALLDIR]\cfg\HALEventTarget-config.x (for UNIX, Linux)

Care must be taken when modifying this file. In particular, the file contains three "blocks" called "params", "target" and "info". The "target" block should not be changed. The "params" block only needs to be changed if you are using a custom script to process events. The "info" block contains one entry for each SNMP manager host receiving traps.

To change the hostname or port of the machine currently setup to receive traps, look for a block named "MgrTrap" inside the "info" block. If there is more than one host receiving traps, find the entry for that host:

info = {
  MgrTrap = {
     host = oldhostname
     port = 162
     tags = event
     params = SnmpTrap
  }
}

and update the values of the "host" and/or "port" tags to point to the new machine:

info = {
  MgrTrap = {
     host = newhostname
     port = 1234
     tags = event
     params = SnmpTrap
  }
}

To add an additional host to the list of trap targets, simply add another block inside the "info" block:

info = {
  MgrTrap = {
     host = hostname_one
     port = 162
     tags = event
     params = SnmpTrap
  }

  MgrTrap2 = {
     host = hostname_two
     port = 162
     tags = event
     params = SnmpTrap
  }
}

The names of the blocks are irrelevant, but they must be unique, and must not contain any "special" characters (punctuation or control characters). There is no limit to the number of trap targets you can add.

For the new configuration to take effect, the agent must be restarted.

31.3 What operating systems and modules are supported with the Neuron Agent?

Since Neuron supports multiple operating systems, the list of modules supported is best summarized in a table, please refer to the following:

Modules	Solaris	Linux	Tru64	Windows
SystemMonitor*	Yes	Yes	Yes	Yes
Agent Health*	Yes	Yes	Yes	Yes
DirectoryMonitor*	Yes	Yes	Yes	No
FileMonitor*	Yes	Yes	Yes	No
HardwareAlert*	Yes	No	Yes	No
LogFileMonitor*	Yes	Yes	Yes	Yes
MIB II*	Yes	Yes	Yes	Yes
ProcessMonitor*	Yes	Yes	Yes	Yes
Reporter*	Yes	Yes	Yes	No
ScriptRunner*	Yes	Yes	Yes	Yes
ScriptPack*	Yes	No	No	No
Cluster Watch	Yes	No	Yes	No
Oracle	Coming Soon	No	No	No
SQL Server	No	No	No	Yes
Sun Directory Server	Yes	No	No	No
Sun Messaging Server	Yes	No	No	No
Sun Web Server	Yes	No	No	No
BEA WebLogic	Yes	No	No	No
Veritas Cluster Server	Yes	No	No	No
Event Gateway	Yes	No	No	No

^*These modules are included in the Neuron Agent package. Other modules are provided as plugins/addons and must be installed and purchased separately. Halcyon is currently in the process of making their modules for Sun Management Center compatible with the Neuron Agent. To view the list of monitoring solutions currently available for Sun Management Center, visit our products page

31.4 What are the requirements for sending SNMP Traps from Neuron to 3rd party managers?

The following is a list of requirements that need to be considered for forwarding events from the Neuron Agent:

Firewall
If there is a firewall between the Neuron Agent and the 3rd party SNMP manager, then the SNMP Trap Handler port (default 162) that is being used by the SNMP manager must be opened up for UDP packets. The HTTP port (default 6680) on which the Neuron Web Server has been installed must also be opened if viewing through a browser that is behind a firewall. Remote acknowledgements (ACKs) will also go through the HTTP port. Note, you do not need to open up the firewall for the port on which the agent was installed (default 6681).



SNMP Support
The machine on which Neuron is to be installed should support SNMP and have available the necessary libraries.

The only seems to apply to the Tru64 OS, where the system must have the following kernel parameters enabled:

X/Open Transport Interface (XTISO, TIMOD, TIRDWR)

To enable on Tru64, add the following lines to the /sys/conf/HOSTNAME file and reboot:
options TIRDWR
options TIMOD
options XTISO

In general, when the standard packages are installed for Solaris, Windows, and Linux, SNMP is usually supported. Once the Neuron Agent is installed, you can manually run the following snmptrap command to verify SNMP traps can be sent out:

# ./snmptrap -h HOSTNAME -p PORT -c public 1.3.6.1.4.1.1242.1 142.9.120.178 6 9 0 1.3.6.1.4.1.1242.1.1.3.6 OctetString Halcyon-Test

Note: ./snmptrap can be found under the INSTALL_DIR/alpha/bin directory on UNIX and C:\Program Files\Halcyon\PrimeAlert\install\bin on Windows.