5.1 Monitoring Server Health

Monitoring the health of your server can help prevent it from getting to a state in which your users cannot access the server or the data on it. Monitoring your server's health involves the following tasks:

5.1.1 Monitoring Overall Server Health or the Health of a Specific Item

Using Novell Remote Manager, you can monitor the server's overall health and the health of specific items on the server.

The server's overall health is indicated by the color of the circle displayed next to the Overall Server Health Status Indicator icon Server's health status is good in the header frame for Novell Remote Manager. The following table lists and explains each health status that might be displayed.

Table 5-1 Overall Server Health Status Explanation

Icon

Server Health Status

Explanation

Good

All parameters included in the server's health configuration list are good.

Suspect

The status of one or more of the parameters included in the server's health configuration list is suspect or has a minor problem.

Bad

The status of one or more of the parameters included in the server's health configuration list is bad or has a critical problem.

Lost connection

The connection to the server from Novell Remote Manager has been lost.

The server's overall health is determined by items that are selected in the Include list on the detailed Health Monitor page. By default, all items are checked.

Therefore, if you have a server that has specific parameters that you know will cause a suspect or bad status and you want to be notified when other parameters have changed, you can remove the suspect or bad parameters from the Include and Notify lists by deselecting them and clicking Apply Settings on the Health Monitor page.

If the status of any item that is selected in the Include list changes to yellow (suspect) or red (bad), the health status indicator light on the header frame changes to indicate there is a problem. If more than one item changes, the worst status indicates the server's overall status. When the status for all items returns to green (good), then the health light indicator changes back to green (good).

The server's health status, reported by the health status indicator, is updated every five seconds, but the graphic refreshes only if the status changes.

You can also view the changes of status in the Server Health Log. For more information, see Viewing and Managing Server Logs.

The Health Monitor page shows the health status (green, good; yellow, suspect; or red, bad) for all known components of the system, as well as current, peak, and maximum values. When an item is not selected in the Include column, it is not included when determining the overall server health and the values for Status, Current, Peak, and Max are not displayed.

Figure 5-1 Health Monitor Page

The following items are key indicators of your server's health:

  • Work To Do Response Time
  • Allocated Server Processes
  • Available Server Processes
  • Abended Thread Count (if the server has abended and recovered automatically)
  • Failed Logins Per Hour
  • CPU Utilization (for each processor, if there is more than one)
  • Faulted Address Space Count
  • Connection Usage
  • Available Memory
  • Available Logical Space
  • Virtual Memory Performance
  • Cache Performance
  • DS Status
  • DS Thread Usage
  • Packet Receive Buffers
  • Available ECBs
  • LAN Traffic
  • Vol Sys Available Disk Space
  • Available Disk Space for all volumes except sys:
  • Available Directory Entries
  • Disk Throughput
  • Support Pack Status

    The status is always green on the Support Pack Status item until Novell is able to provide Support Packs in a format that is consumable by this application. When that occurs, the options provided by this status and the information page explaining how to consume this information will be updated.

You can change the thresholds for the Suspect and Critical values of these indicators individually or restore the defaults for all these parameters at one time. See Configuring Health Thresholds.

Additional items are tracked on the Health Monitor page and are displayed depending on the NLM™ programs that are loaded on the server.

To access the Health Monitor page, click one of the following links in Novell Remote Manager:

  • Server's health status is good Overall server health status indicator icon.
  • Health Monitor link in the header frame Health Monitor icon in the header frame
  • Health Monitor link in the navigation frame under the Diagnose Server heading.

5.1.2 Configuring Health Thresholds

If you have a server that runs differently than normal, you can configure the Suspect and Critical threshold values for Health Thresholds to something other than the default or change the amount of time before the trigger occurs. After changing setting, you can also restore all settings to the default values. For procedures, see the following:

Changing Criteria for Suspect or Critical Health Indicators

To change the criteria for the suspect (yellow/suspect) and critical (red/bad), status indicators for a specific operation:

  1. On the Health Monitor page, click Threshold Configuration, then click the Description name link for the item you want to change.

  2. Specify the new Suspect or Critical value, then click OK.

  3. (Optional) If you want to change both the values back to the default setting, click Restore Default Values, then click OK.

Changing Suspect or Critical Status Indicator Triggers

To change the amount of time before a status change for the suspect (yellow/suspect) and critical (red/bad) status indicators for a specific operation:

  1. On the Health Monitor page, click Threshold Configuration, then click the Description name link for the item you want to change.

  2. Specify the new Suspect or Critical trigger delay value, then click OK.

  3. (Optional) If you want to change both the values back to the default setting, click Restore Default Values, then click OK.

Restoring Settings

To restore all the settings to the default values after changing them, click Health Monitor > Threshold Configuration > Restore All Defaults.

5.1.3 Viewing Statistics Trend Graphs

The Statistics Trend Graphs let you select and create specific server statistics trend graphs for the statistics and duration listed. The graphs can cover from the last minute up to the total time the system has been up, depending on the duration you select.

After changing the server configuration or hardware, you can reset the data you gather.

All the graph data is saved to files in sys:system with a .tdf extension. These files use minimal disk space compared to the value they provide.

For information about the trends you notice and actions to take, see the online help.

Table 5-2 Tasks and Procedures for Viewing and Generating Server Statistics Trend Graphs

To

Do This

View the health statistics trends for your server

Access the Statistic Trend Graph page by clicking the Diagnose Server > Health Monitor or the Manage Server > View Statistics link in the navigation frame and then clicking the Statistics Trend Graphs link on either of the displayed pages.

Determine which graphs you want to view and the duration for the trend

On the Statistic Trend Graph page, select the check box for each statistic and duration you want to view, then click Draw Selected Graphs.

View one or more graphs

On the Statistic Trend Graph page, select the check box for each statistic and duration you want to view, then click Draw Selected Graphs.

A graph is drawn for each check box you select.

If you select durations of an hour or more, an average as well as a maximum value for each statistic is shown. The maximum value is always the top line of a graph.

If you select durations of a week or a month, each graph also has links to display statistics for specific days or weeks depending on the duration you first selected.

The All Data column shows data for each statistic since portal.nlm was started on your server or the Reset button was clicked for a specific item. If the server is shut down, the statistics are still tracked and recorded as 0.

Clear all the check boxes and start determining what you want to view again

On the Statistic Trend Graph page, click the Reset button at the bottom of the selection table.

Reset the data for a specific statistic

On the Statistic Trend Graph page, click the Reset button at the end of the row for that statistic.

Reset the data for all statistics

On the Statistic Trend Graph page, click the Reset button for each of the statistics.

Draw your own customized graph

On the Statistic Trend Graph page, view the raw data for the graph by clicking the View Hourly Raw Data link, then copy and paste the numbers into a spreadsheet/graphing application.

5.1.4 Configuring E-Mail Notification for Server Health Status

Rather than manually checking the server’s health status through Novell Remote Manager’s interface, you can configure Novell Remote Manager to send an e-mail to notify you when the server's health status changes to any value other than Green (good).

  1. On the Health Monitor page, click the Mail Notification Configuration link.

  2. Specify the required information on the Mail Notification Configuration page, then click Save.

    A primary server and at least one user on the notification list is required.

    If you want to configure the same information on several servers, create a group using the Group Operation procedures and then run the Synchronize E-mail Settings operation. For more information, see Building and Configuring a Monitoring Group and Performing Synchronizing Operations on Server Groups.

  3. On the Health Monitor page, select the Notify check box for each item you want notification for.

    If you want to receive an abend log by e-mail when the server has abended, click the Notify check box for the Abended Thread Count item on the Health Monitor page.

    The e‑mail contains a default subject line: Server Health Warning.

    If you want the subject line of the e-mail notification to contain information about the server, the specific health item the notification is for, and its status, do the following:

    1. In the header frame of Novell Remote Manager, click the Configure icon .

    2. On the Novell Remote Manager Configuration Options page, click the Email Detailed Subject Line for Health Alerts option to toggle the status to Yes.

      Clicking the Email Detailed Subject Line for Health Alerts to toggle the status to No returns the subject line for the e-mail notification to the default subject line.

5.1.5 Troubleshooting a Suspect or Bad Health Status

When the health status of an item changes from good to a suspect or bad state, you can look at the specific item and click the Info icon Info icon for suggested remedies or click the link in the Description column for more detailed status and information about the health item.

  1. Access the Health Monitor page.

  2. Look for the specific health item that has changed status.

  3. View the information for the item that has changed by clicking the Info icon Info icon for the item.

    This information outlines the specific health criteria (thresholds) for green, yellow, red or statuses in that component. It also provides suggestions in some cases for what might be going wrong in that component if a yellow or red indicator is displayed.

  4. Perform the recommended or appropriate action for the health item that has changed.

  5. In the Description column of the Server Health Table, you can also click the link for the applicable item to view more detail information tracked about the health item.

You can also use the Server Health Log to track the status of each health item that changes. For more information, see Viewing and Managing Server Logs.