NetVault SmartDisk 11.4.5 - Administration Guide

Using resource management

Based on how you configure it, the resourcemanager can check the trace files every minute after NetVault SmartDisk starts. The resourcemanager reads the configured values at the start of every repeat cycle, for example, every minute, which ensures that it reads configuration changes that are made while the process is running.

Using the Modules section in the /foundation/etc/resourcemanager.cfg file and the ResourceManagement:TraceDelete section in each module’s /etc/trace.cfg file, you can specify which modules you want the resourcemanager to monitor and manage.

In the following example, the resourcemanager is set to manage the diskmanager, dedupe, and foundation modules.

NetVault SmartDisk creates several types of trace file, *.trc, *.buffer, *monitorrecent*.txt, and *monitortrace*.txt, but only certain ones are eligible for deletion by the resourcemanager. Understanding the naming convention applied to these files is critical, because it is key to how the resourcemanager determines what to include or exclude from its eligible list. The following table identifies the naming convention used for the different files and the different platforms.

The resourcemanager only considers trace files as contributing to the totals it compares against the start and stop thresholds if it identifies them as inactive; if the files are inactive, it indicates that they are not associated with a process that is running. It determines this state by comparing the process ID in the trace filename with the list of IDs for processes that are running on the system at the start of each repeat cycle. Because the process ID is not used in the filename on all operating systems, the resourcemanager can delete *.trc, *.buffer, monitorrecent*, and monitortrace* files on Linux, UNIX, and Mac OS X, but it can only delete *.trc and *.buffer files on Windows, as shown in the preceding table.

Specifying which modules to manage

NOTE: You can specify any module. The resourcemanager only applies trace deletion to the specified modules if it locates them in the directory in which NetVault SmartDisk is installed. If it cannot locate a specified module, it generates its own error-level trace message, but continues to function normally.

Configuring the Trace-Delete option in specific modules

To limit the overall number or size, or both, of trace files in an installed module's trace directory, you can configure the resourcemanager to delete them between start and stop number and size thresholds. You can set these thresholds through the ResourceManagement:TraceDelete section of a module's etc/trace.cfg file. If that section is not present or is invalid, hard-coded defaults are used.

The ResourceManagement:TraceDelete section in the <moduleName>/etc/trace.cfg file lets the resourcemanager determine which trace-delete parameters to apply in the specified module. The defaults are:

If you change the values and inadvertently render them unacceptable, not present or invalid, the hard-coded defaults are used instead.

In the preceding example, which shows the defaults, the Trace-Delete functionality is enabled, and the resourcemanager starts deleting the trace files, oldest first, that exist in the specified module’s /trace directory when their combined number becomes equal to or greater than 100 or their combined size becomes equal to or greater than 1 GB. The files continue to be deleted until their combined number is reduced to 10 or their combined size reaches or exceeds 1 MB.

To modify the values, perform the following steps.

Disabling trace deletion

You can disable trace deletion for all modules or on an individual basis.