FSRM-enabled disk takes a long time to come online in Windows Server

Symptoms

Assume that you have a File Server Resource Manager (FSRM) feature-enabled volume that's a cluster disk. This disk contains a large volume of data and many folders that have quota settings applied. When a cluster failover occurs in this situation, the cluster service tries to bring the disk online on another node. However, the attempt is unsuccessful.

Cause

This issue occurs because an FSRM quota database full rescan is triggered when the failover occurs. The full rescan scans all the data to verify and rebuild the quota database. Depending on the data size, the number of folders that have the quota setting applied, and the disk storage performance, the time that's required to complete these tasks may exceed the default cluster disk online pending time-out. Therefore, the attempt to bring the disk online fails.

Workaround

To prevent this issue, use one of the following workarounds: 
  • Increase the disk online pending time-out to a value that's more than the FSRM disk online time.
  • Avoid putting a large volume of data on a single volume when FSRM is enabled.
Note Because disk and file system I/O performance varies on different systems, even with the same data size and quota settings, there is no static limit on data size or quota settings. These details should be tuned under specific hardware and software configuration scenarios.

More Information

When this issue occurs, the following event, showing a quota full rescan, is triggered:


When the cluster disk online failure occurs, and event that resembles the following is triggered:

Properties

Article ID: 3053651 - Last Review: 20 Apr 2015 - Revision: 1

Feedback