The bug is in a Microsoft Update scan engine that debilitates performance when a client is undergoing a routine scan. The fix is expected to be available by the end of September.
The problem occurs with the Systems Management Server 2003 SP1 or later using the Inventory Tool for Microsoft Updates (ITMU) and on machines that are running the Windows Update agent. It can cause Microsoft Update and Windows Server Update Services (WSUS) to take up to five minutes per scan on a client.
As Microsoft releases updates, the number of files in the scan tool grows. Additional files in the tool require that you have more processing power to continue to scan each system for updates. "If [administrators] haven't had the problem at this point, it's because they have fast computers or their end users do have the problem and they just don't know about it," said John Juris, a former SMS administrator who is now a partner at SMS Expert, a Chicago-based company that makes tools to use with SMS.
The delays could increase the number of calls to corporate help desks, and that could have an impact on all businesses, Juris said.
The ITMU tool is designed to help customers update their managed systems by replacing a multitude of legacy scan tools with one scan tool, Microsoft said. The tool shares the same security update, update rollup and service pack data as offered by WSUS.
This inventory tool can be used only within an SMS 2003 SP1 site hierarchy with certain hotfixes applied, according to Bill Corrigan, director of marketing for the Windows Enterprise Management division at Microsoft.
Corrigan said Microsoft is working on a resolution for its customers and expects to have a KnowledgeBase article ready within 30 days, though he didn't say exactly what form that resolution would take.
Until then, Microsoft recommends that customers take the following steps to alleviate the problem:
For managed clients, schedule updates outside of business hours. Microsoft is recommending that customers let their end users know that performance may be affected during a specified period of time. Corrigan said to encourage end users not to turn their computers off if they experience a performance issue.
For managed servers, schedule updates at times when server utilization is low.
For ongoing health checks, where applicable, reduce your scan intervals to once a week.