Cluster Bottleneck Occurs When a Large Number of Disks Are Involved

This article was previously published under Q234544
This article has been archived. It is offered "as is" and will no longer be updated.
"High-end" computer systems (especially large FibreChannel installations) that have a large number of hard disks may experience slow response times.
This is because of the number of times the bus is reset. Each reset process can stop the entire bus for seconds at a time. When a large number of hard disks are involved, this time can grow to minutes.
This problem was fixed by reducing the number of bus resets performed, while maintaining the requirement of finding the correct disk device when needed.

To resolve this problem, obtain the latest service pack for Windows NT 4.0. For additional information, click the following article number to view the article in theMicrosoft Knowledge Base:
152734 How to Obtain the Latest Windows NT 4.0 Service Pack
The latest service pack is also available at the following Internet location:
Microsoft has confirmed that this is a problem in Windows NT 4.0. This problem was first corrected in Windows NT 4.0 Service Pack 5.

Article ID: 234544 - Last Review: 11/02/2013 22:30:00 - Revision: 3.0

Microsoft Windows NT Server 4.0 Standard Edition, Microsoft Windows NT Server 4.0 Enterprise Edition

  • kbnosurvey kbarchive kbbug kbfix kbqfe KB234544
ERROR: at System.Diagnostics.Process.Kill() at Microsoft.Support.SEOInfrastructureService.PhantomJS.PhantomJSRunner.WaitForExit(Process process, Int32 waitTime, StringBuilder dataBuilder, Boolean isTotalProcessTimeout)