Improvements in Inventory Solution from 7.0 to 7.0 SP1

Out of all the Solutions updated by Client Management Suite 7.0 SP1, Inventory Solution had the most significant changes. This document outlines what changes were made between the two versions, and what the changes entail. To summarize, the changes make working within Inventory much more intuitive and easy to manage. Out of the box we now more aggressively obtain the Inventory needed, eliminating the need to work through multiple tasks to obtain all the data.

Introduction

For those who used or evaluated Inventory Solution 7.0, you may have found the process of capturing a full inventory confusing and cumbersome. In SP1 the Inventory team strived to fix this by making Inventory much easier to configure and use. To understand why the 7.0 version had the cumbersome process it did, it helps to understand the history of issues stemming from Inventory Solution.

When Inventory sent a full Inventory two thirds or more of Inventory was detailed file data from the audit scan. This caused more network bandwidth utilization and a higher load on the Notification Server. To rectify this, in the original release of 7.0 the file details were split into two types, the basic and the detailed. Due to how this was setup, it would take a full week if nothing was changed out of the box for the Notification Server to have the full file details from any system.

In 7.0 SP1 the default polices send all the file data. While this does again present the potential for Network Bandwidth and Notification Server load issues, there are mechanisms in place to avoid it.

Inventory Policies

Inventory Solution has moved to a primarily Policy-based approach. This allows us to gather Inventory regardless if the machine is available on the network or not. This fits the use-case model for Inventory better than a Task structure. The policy has had its complexity reduced to make it simpler to use, though we have kept the ability to customize it fully.

Scheduling

Normally we might talk about scheduling towards the end of an item as the configuration of the item usually comes first. With the changes in SP1 scheduling is the first option presented to you. To access an Inventory Policy in the Symantec Management Console go to Manage > Policies > Discovery and Inventory > Inventory. The following screenshot shows the configuration of a Policy:

Instead of forcing users to look at scheduling options directly, the model has been switched to a Compliance model. You'll notice under the Applies To section there is a bar for Compliance. The compliance check can be set to Daily, Weekly, Monthly, or a custom schedule can be defined. How this works is the compliance check will see if the computer has sent Inventory within the last day, week, or month, depending on which option is selected. If it has not, it is considered non-compliant and it will run the Inventory Policy immediately. The actual schedule to run the compliance check is as follows:

• Daily: Compliance check runs at 6pm every day
• Weekly: Compliance check runs on Monday at 6pm every week
• Monthly: Compliance check runs on the 1st of the month at 6pm every month

The Compliance bar is based off of how old or new Inventory data is for all systems within the Target. The following screenshot shows a bar with a higher sampling of target systems to illustrate how this bar looks:

The following describes what data makes up this bar:

1. N Computers - Count of all targeted computers - only those computers where the corresponding policy has currently been applied to. Length of progress bar represents the count of computers.
2. N% Compliant - All targeted computers where inventory succeeded within the specified schedule. Progress bar color is green.
3. N% Critical - All targeted computers where inventory succeeded at some point in the past, but now are out of compliance, i.e. inventory data is stale. Scenarios such as computers who earlier where reporting inventory but now are powered off, or having the Altiris Agent stopped or disabled. Progress bar color is yellow.
   NOTE: The targeted computers where the Inventory plug-ins has been uninstalled should not be included in this status, nor in any other status.
4. N% Failed - All targeted computers where Inventory plug-in is installed, the policy started executing but then it failed. Progress bar color is red.
5. N% Unknown - All targeted computers where Inventory task is currently running or is waiting for the previous task to finish before it can start. Progress bar color is white (empty progress bar)

One more note: unless you use a custom schedule, these policies are automatically marked as Run Once As Soon As Possible (ASAP) so that new systems are put into compliance immediately. This ensures new systems are left without Inventory until the next scheduled compliance check.

Maintenance Windows

One item to point out is how Inventory interacts with Maintenance Windows. Inventory does not adhere to Maintenance Windows by default. In 7.0 Inventory would run full throttle inside a maintenance window, but if the maintenance window closed during the run, it would throttle back, usually taking a long time to complete. Due to feedback on this issue, Inventory will no longer throttle automatically outside of maintenance windows. You can toggle whether to adhere to Maintenance Windows if you choose the custom schedule option.

Data Classes

As you may see from the screenshot, the selection of what type of Inventory to collect boils down to 4 checkboxes. To clarify what these mean, see the descriptions below.

• Hardware and operating system - This set of data classes includes the bulk of Inventory's data classes. Only the Software related ones are not included in this option.
• Software - To clarify, this is Software Discovery, which uses the Software Management Framework agent to capture Add/Remove Programs details. Inventory makes direct calls to the SMFagent.dll to gather this data.
• File properties - This is the audit scan that scans all files on a system and reports application or other files based on custom configuration.
• Server applications - This only applies to servers, and if you have Inventory Pack for Servers installed and licensed. It will be grayed out if this Pack is not installed.

As you can see, the choices are extremely slim. For most people, this is sufficient. If you need to be more granular on what data classes you want captured, click the Advanced button and see the next section: Advanced Options.

Advanced Options

There are three tabs under the Advanced options, detailed below.

Data classes

In the original 7.0 release an administrator was presented with a tree of data classes with check boxes. This tab reveals this same interface and allows you to be granular on what data classes you want captured. The tree is split first into data class type, and then into platform (Windows, Mac, etc), and finally into individual data classes. See this screenshot for an example:

File Properties Scan Settings

The settings under this tab include inclusion and exclusion rules for files, folders, and drives, and the specific file rules for what types of files are captured. In the 6.x version of Inventory Solution this functionality was held by the AeXAPEdit.exe utility. In 7.0 to 7.0 SP1 these options have not changed.

Run Options

The Run Options are important in controlling how Inventory is captured. One crucial option has been added in SP1 to give a huge degree of ease of use. The following options are worth calling out:

1. Send inventory changes (deltas) only - My recommendation is to send a non-delta Inventory once a week. This helps keep the client and server in sync.
2. Process priority settings - If you are experiencing disk I/O impacts, set the priority to low or use the override option and set it to 50 > 30 respectively. From there work down until you find the middle ground between resource impacts and how long the Inventory Policy takes.
3. Distribute sending inventory evenly over N hours - This option will help alleviate a problem that caused us to originally split the file inventory into two processes, basic and Detailed. If you schedule a full Inventory on Monday at 6pm, for example, and you have a huge hit on the network and the NS resources when people boot their computers the next morning, provide a value here that will stagger how quickly Inventory comes in from the target systems. For example 12 hours will give ample room for the network and the NS to handle incoming Inventory.

See this screenshot for an example of this window:

Inventory Tasks

Inventory Tasks are no longer used by default in Inventory Solution. Due to the nature of Inventory, and the theory behind regular updates by Inventory Capture, The Task mechanisms don't fit well into the use model. However Tasks are still available as there may be less used use-cases that warrant having the ability to push an Inventory Capture out to a target of systems.

If doing an upgrade, the previous Tasks created by the default install of 7.0 or custom created by an administrator will still exist under a folder labeled Legacy Tasks. See the screenshot under the next section labeled Task Server for a view of the tree.

Task Server

Task Server is a push methodology for executing a job or task. This means that for all the targets in an Inventory Task, a Task Server has to be available and able to reach the target systems at the time of execution. This is especially problematic for systems that are shutdown, or laptops which are frequently disconnected. This is the primary reason Inventory has moved to a policy-based method. As I mentioned previously, however, there may be use cases where having the ability to push out an Inventory Task via Task Server is warranted. Tasks are still available.

The following screenshot shows the changes that have been made to a Capture Inventory task:

Scheduling

When scheduling a task the standard scheduling for Task Server is used, including the following run options:

• Quick Run - a computer can be selected individually
• New Schedule:
  • Now - Allows the Task to be pushed down to the clients immediately upon save of the schedule
  • Schedule: - Allows a scheduled time to be set for the task to be executed. This will still require a Task Server to push down the task
• Override Maintenance Windows - This checkbox allows an administrator to allow the Task to be run outside of defined maintenance windows. Please see the previous section under Policies for Maintenance Windows for more information.

Again these options are enforced by Task Server. For example if you schedule an execution of Inventory for tomorrow, the clients won't know about it until the scheduled time as it is the Task Servers that initiate the task in a push method.

Licensing

In the original 7.0 release Licensing reacted differently to a computer's status than it did in version 6.1. This created a problem for those who upgraded and found that their license count was now being figured differently, putting them out of license compliance. This KB details the differences between the two models:

• https://kb.altiris.com/article.asp?article=48810&p=1

In SP1 licensing now works as it did in 6.1, allowing administrators to upgrade without having to refigure how they manage the statuses of their systems and putting them back in license compliance.

Stand-alone Inventory

In SP1 the Stand-alone package builder now has the same look and feel as an Inventory Task or Policy, with the same data class options. Initially they are presented as the four check boxes detailed earlier, with an Advanced button to give more granular control to the user if they needed it.

Conclusion

The improvements to Inventory Solution in SP1 should allow a much more seamless process, where Inventory is collected promptly. The compliance structure allows a quick review of Inventory compliance, enabling an administrator to know if the report data he or she is reviewing is accurate and up to date.