Baseline automation taking longer than manual process

Content Authoring
1

(imported topic written by RichB91)

I have a series of 9 Tasks assembled in to a Baseline.

At a very high level the 9 tasks do the following:

  1. Verify the existence of, and create if missing, three directories

  2. Read a value from the registry, write that value to a temporary text file, set the registry value to zero

  3. Use Pause Until to give a program up to 15 seconds to close gracefully before proceeding

  4. Issue a restart with a 15 second delay

  5. Verify the registry settings and that a program that shouldn’t be running, isn’t running

  6. Run a dummy-program to simulate our patching cycle for testing

  7. Read the value from the temp file in step 2 and set the registry key to that value

  8. Issue a restart with a 15 second delay

  9. Verify the registry setting and that the program that should be running, is running within 120 seconds of recovering after the reboot

Each step appears to work as intended, both individually and as a baseline. The problem is with how long that Baseline takes to complete.

Averaged across three tests, the

Baseline

took

29 minutes 4 seconds

to complete these 9 steps. Averaged across two tests, doing these steps

manually

took

14 minutes 28 seconds

.

In an effort to compare apples-to-apples, when done manually, I counted off the same 15 and 120 second delays that should happen in the Baseline. But the manual process also meant running RegEdit and browsing to the right location, running Notepad including browsing to the correct folder and naming the file to be saved, actually logging in to the test machine twice rather than just rebooting, and several other minor steps that were not required by the Baseline.

And I think it is safe to say that if I had run this as some kind of VBScript locally that it would have been even faster.

Based on the time stamps in our audit trail, I suspect that the “extra†time is less in the actual individual Tasks and much more in the gap-time between one Task in the Baseline finishing and the next Task starting.

Any ideas why the Baseline would take twice as long as the manual approach?

Do you think modifying any of the following have a significant beneficial effect?

• _BESClient_Report_MinimumInterval

• _BESClient_Report_MinimumAnalysisInterval

• _BESClient_Comm_CommandPollIntervalSeconds

• _BESClient_Resource_WorkIdle

• _BESClient_Resource_SleepIdle

2

(imported comment written by MattBoyd)

Baselines seem to complete faster for us after changing _BESClient_Resource_WorkIdle and _BESClient_Resource_SleepIdle to utilize the CPU more. I don’t think the other settings that you listed will improve the performance of baselines, and could have additional consequences.

There’s a BigFix task in the BES Support site called “BES Client Setting: CPU Usage” that you can use to do this. If you want the settings to take effect immediately, restart the BES client after applying the CPU usage task.

Also, I I’m not really sure what you’re trying to accomplish with this baseline? Can any of these tasks be combined to reduce the number of tasks in the baseline?