In certain situations when a Windows 2000 Server-based file
server is under a heavy load or stress, the NTFS log file may not flush the
disk in time, and as a result becomes full. When this behavior occurs, NTFS file
operations do not continue until the log file is completely flushed to disk.
You may experience decreased performance during this time.
If you
experience decreased performance on a Windows 2000 Server-based file server and
you want to determine whether it is caused by a bottleneck in the NTFS log file
size, in
Performance Logs and Alerts, view the activity of the
Current Disk Queue Length counter for the appropriate
PhysicalDisk object.
If the performance data shows constant high disk queue length values that intermittently drop to a value of exactly 1 for a period of time (during this period you cannot write to disk) and then increases again, this indicates that the NTFS log file is full and must be flushed.
Note When the value of the disk queue length drops to one, you cannot
write to disk.
To troubleshoot and resolve this issue, follow these
steps:
- Increase the performance of the disk subsystem. For
example:
- Install additional disks, or upgrade the existing hard
disks. Update the bus and disk controllers.
- Use striped volumes on several physical disks to
increase throughput.
- Increase the NTFS log file size. To do so, follow these
steps:
- Click Start, click
Run, in the Open box, type
cmd, and then click OK.
- At the command prompt, type the following line, and
then press ENTER, where Size is the size of the log
file:
chkdsk /l:size
If stress on the disk subsystem continues to be high, the log file
may eventually become full again. Use this method if stress on the disk system
is temporary.
For more information, click the following article number to view the article in the Microsoft Knowledge Base:
101670
Transaction log supports NTFS recoverability