This Knowledgebase article provides information about Neverfail Heartbeat V5.3 Build 1804 - Beta Release.
A single unified Application Management Framework (AMF) replaces the separate Application Manager and Application Monitor components available in previous versions of Neverfail Heartbeat providing:
- Management of all aspects of protected applications with in the Management Client.
- Integrated protection of application services without the need for start / stop scripts.
- Passive Server service monitoring prevents protected applications from starting on the passive server – this prevents most file synchronization problems.
- Plug-ins can be upgraded without stopping Neverfail Heartbeat.
- Backwards compatible with existing AM(X)s installed using SETUP.
- Introduction of 'Plug-Ins' as components, which constantly adapt to changes in the application configuration ensuring dynamic application protection. This extends to file filters, services, tasks, and monitoring rules, which are also integrated into the Management Client.
- Regardless of Heartbeat server version, all server pairs can now be administered from a single Management Client.
- Users can now define 'Exclusion' file filters to exclude content from the protected set.
- Neverfail Heartbeat can now be stopped from the System Tray icon.
- Performance improvements.
- Added the ability to choose windows or java look and feel with Windows PLAF as the default.
- The Reference Guide and Installation Checklist have been removed from the installation kit. Information about technical documentation can be found in the Readme.txt file contained in the kit. All documentation is now available from the Neverfail Extranet.
- Temp files created during Full System Check are now compressed.
- Incorporated SCOPE Data Collector service V4.2.0.
- Removed the case sensitivity of the \Windows folder on the Primary and Secondary server, which previously resulted in "The root folder is different" error message during Setup.
- Removed "while active/passive" from the default red alert, yellow alert email and tool tray messages.
- Removed the 'Mark as In-Sync' and 'Mark full file system as In-Sync' buttons.
- Changed Red/Yellow Alerts so that they are logged to event log. Added a new EventLogResponse. Removed the default
- Red/Yellow exec command. Added ability to set alert event log message and enable/disable logging.
- Changed "Error sending email alert (KB 517)" Alert. Modified the EmailResponse class to fire EmailFailedEvents only when the previous attempt to send was successful.
- The Server Configuration application now uses the same look and feel as the Management Client.
- System Status and Control Panel changed to provide more detailed information about the health of the system.
- Redesigned to indicate whether the Primary or Secondary server is providing service
- Indicates whether applications are running or stopped.
- Health status (Ok or Degraded) of an application (either because services are not running, or indicated by triggered rule).
- Applications can now be started and stopped from this panel.
- Application monitoring can be temporarily disabled for the purpose of maintenance. This is performed via the 'Configure' dialog.
- Timeouts can be set via 'Edit'.
- The services panel indicates whether services have been protected by an AM(X)s or by the user.
- Service dependency is also indicated in the GUI as these are monitored as well. The target state, of protected services, can be specified for the active and passive servers.
- Services depending on protected services will be managed (i.e. started and stopped) by Heartbeat but not monitored (i.e. not restarted if stopped by some external agency).
- Services upon which protected services depend will be monitored (i.e. restarted if stopped) but not managed (i.e. not stopped if protected applications are stopped).
Tasks -these are an extension of the start, stop, and monitor scripts in earlier versions of Heartbeat. Types of task are classified by when the tasks run, and include the following tasks:
- Network Configuration – These types of task are run first to ensure the environment is updated to support the application. Typically used in a WAN, they are used to launch dnscmd and dnsupdate. Where multiple dnscmds are required, these can be contained in a batch script, which is launched by the task.
- Pre/Post Start - Run before and after services are started on the active server.
- Pre/Post Stop - Run before and after services are stopped on the active server.
- Pre/Post Shadow -Run before and after a shadow copy is created on the active server by the Data Rollback Module.
- Rule Action - Run only when a rule has been triggered, or a service fails its check.
Rules - these are an extension of the current application and system monitoring rules. All rules can only implemented by plug-Ins (i.e., there are no user-defined rules).
- Rules can be either “timed” (they must evaluate as true continuously for the specified duration in order to trigger) or “latched” (they trigger as soon as they evaluate to true).
- Rules can be configured with rule actions, which are the tasks to perform when the rule triggers.
Plug-Ins - Plug-Ins are .NET Assembly .dlls that implement all of the functionality required to provide application-specific protection for Heartbeat.
- Plug-Ins are installed by browsing to the *NFPlugin.dll file from Heartbeat.
- Plug-Ins can provide configuration options.
- Summary panel
- Filters now have one or more 'Sponsors'. A protected application provides sponsorship of a filter, to indicate it requires this filter to be present. Filters can also be sponsored by the user.
- Exclusion filters - These must be subsets of (i.e., be more restrictive than) inclusion filters.
- Filters are compared with each other, and if one filter is more general than another is, only the more general is used. This allows efficient filter sets to be used, e.g., fileservers with 1000s of individual shares requested by an AM(X) or Plug-In can be replaced by a single more general filter configured by the user.
- All filters are validated before being passed to the replication components, and the GUI reports why filters have been rejected (e.g., because the drive does not exist).
- Configuration panel
- Corrected a potential NullPointerException if the exit was called before initialization.
- Corrected the FullSystemCheckTask so that the processor does not sit at 100% for a 15-minute period when verifying a 60GB file with LBM enabled.(ID 5660)
- Removal of hard coded foreground color settings. (ID 5691)
- Corrected to only report errors on editingStopped() if the editor contains a non-empty string. Also, editingStopped() now ensures that editor's text is cleared before returning. (ID 5581)
- Corrected a handle leak when FSM attempts to detect reparse points. (ID 5697)
- Corrected issues installing the nfpktfltr during Setup on Windows 2000 installations when the Digital Signature Not Found dialog fails to appear. (ID 5589)
- Corrected NFServerR2: InvokeJVM - Failed to convert Java Arg list. Problem installing on paths with chars from multiple code pages. (ID 5348)
- Corrected the display, age of certain entries were appearing in the wrong queue. Safe queue entries displayed in the Unsafe queue and vice versa. (ID 5640)
- Corrected the application status (blank) on the Secondary server when the GUI is restarted and the channel is not connected. (ID 5541)
The following are known issues in Neverfail Heartbeat V5.3.0 build 1804 Beta.
Environment and Configuration Limitations
In Heartbeat V5.3.0, the existing Application Modules are superseded by Application Framework Plug-ins.
Only the following plug-ins should be used:
- Microsoft Exchange (for Exchange 2003 / 2007)
- Microsoft SQL Server (for SQL Server 2000 SP4 / 2005)
- File Server (on Windows 2000 / 2003-x86 / 2003-x64)
- None of the existing Application Modules should be installed with this release of Heartbeat.
Neverfail Heartbeat should only be installed on English-language versions of Windows.
- In Heartbeat V5.3.0, the existing Application Modules are superseded by Application Framework Plug-ins.
Application Framework Plug-ins cannot be installed successfully via Setup, because the necessary files are not copied to the correct location.
- On both servers, unzip the Neverfail Heartbeat installation kit.
- On both servers, unzip the plug-in kit.
- Install Neverfail Heartbeat on both servers. On the Primary server, when asked if you wish to install an AM(X), select the plug-in you wish to install.
- After installation completes, ensure Neverfail Heartbeat is shut down on both servers
- On the Primary server, open the folder containing the Neverfail Heartbeat Installation kit. This will contain a folder with a name corresponding to the plug-in you installed e.g. 'SQLServer'. Copy the folder and all its contents into the location where you installed Neverfail Heartbeat, under R2\plugins. For example, if your installation kit is at C:\temp\V5-3-1804-X86, and Neverfail for SQL Server was installed into C:\Neverfail, then you would copy C:\temp\V5-3-1804-X86\SQLServer to C:\Neverfail\R2\plugins\SQLServer.
- Copy the same folder and its contents to the same location on the Secondary server.
- Start Neverfail Heartbeat on both servers.
It is possible to install Neverfail Heartbeat, and subsequently add the File Server Plug-in. If this is done, filter discovery for the File Server Plug-in will not run.
- Open the Neverfail Heartbeat Management Client.
- Navigate to Application -> Status .
- Click 'Stop Applications'
- Click 'Start Applications'
Upgrades from earlier versions of Neverfail Heartbeat should not be attempted, as the upgrade functionality is not fully implemented yet.
- Application Framework Plug-ins cannot be installed successfully via Setup, because the necessary files are not copied to the correct location.
- The Neverfail Heartbeat Management Client may stop responding if your Neverfail Heartbeat server is under a heavy load.
- Occasionally, the Neverfail Server R2 service may exit spontaneously, issuing a 'Neverfail Server R2 service terminated unexpectedly' message in the Windows System Event log.
Occasionally, a switchover operation may hang.
- Shut down Neverfail Heartbeat.
- Reconfigure the server roles as they were prior to the switchover.
- Reboot the server pair. This will also start Neverfail Heartbeat on both servers.
- Wait for the Full System Check to complete.
- Attempt the switchover again.
- User-defined tasks created while one of the servers in the pair is active, will not be available following a switchover, when the other server becomes active.
Tasks which are deleted from the
section of the Neverfail Heartbeat Management Client will continue to run. Following a switchover, the deleted tasks will be reinstated, and will reappear in the Management Client.
- After deleting one or more tasks, shut down Heartbeat on both servers in the pair, and then start it again.
- Start and stop timeouts for protected applications may be set through the Management Client, via Application -> Status , by selecting an application and clicking the 'Edit' button. These changes are persisted, but they do not actually modify the application's behavior.
Following a failover, the failed server is automatically restarted as passive. This is by design, in order that the protected application services are not launched on the failed server, and also to prevent this server attempting to connect to the network. With Heartbeat V5.3.0, following a failover, the failed server is still recovered as passive, and will still be disconnected from the network; but the protected application services may nevertheless be started.
- Shut down Heartbeat on the active server.
- Check that Heartbeat on the failed server is in the passive role.
- Start Heartbeat on the passive server.
- Wait for 2 - 3 minutes, and then start Heartbeat on the active server.
When Heartbeat is started on the passive server, the protected application services may be started, and their startup type set to 'Automatic'. Neverfail Heartbeat will detect this condition after about two minutes, and shut down the services, setting their startup type back to 'Manual'.
- When starting Neverfail Heartbeat, start the passive server first, and wait for at least two minutes before starting the active server. This reduces the risk that application services will be running on the passive server during a Full System Check.
- If you attempt to protect the root of a drive, the Application Framework will not automatically veto the filter. Setting such a filter is not recommended even for current versions of Neverfail Heartbeat: see Neverfail Knowledgebase article #399 - 'Attempting to Protect the Root of Any Microsoft Windows Drive Using a File Filter (e.g. C:\** or D:\**) Is Not Recommended.
- Setup is unable to uninstall NFServerR2.exe, the executable for the Neverfail Server R2 service.
- On uninstallation, you may be asked to delete the root folder of the Neverfail Heartbeat installation (by default, C:\Neverfail). This folder is in use by Neverfail SCOPE, which is not removed when Heartbeat uninstalls; so you should ignore the warning and leave the folder in place.
If a plug-in is uninstalled, a subsequent attempt to reinstall the same plug-in may fail.
On BOTH servers:
- Under the Heartbeat installation directory, navigate to \R2\plugins.
- Delete the directory corresponding to the Plug-in you want to reinstall.
- Reinstall the plug-in.
Technical Documents supporting this version of Neverfail Heartbeat are contained in Knowledgebase article #1280 - 'V5.3[.n] Neverfail Heartbeat Technical Documentation'.
Installation instructions are found in Neverfail Heartbeat and Neverfail Replicator - Technical Documentation - Reference Guide V5.3 .
Neverfail Heartbeat V5.3 Build 1804 - Beta Release