To troubleshoot this issue, run the ChkNTFS utility to
determine the state of the dirty bit. If the bit is reported as dirty, see the
following article for additional troubleshooting steps:
176970 How to Run the CHKDSK /F Command on a Shared Cluster Disk
If the bit is not reported as dirty, this issue may
be caused by a permissions issue. When the Cluster service brings a physical
disk resource online, it performs the following steps to detect potential
corruption of the disk resource:
- It opens a temporary file named
zClusterOnlineChk.tmp at the root of the hard disk.
- It writes a test pattern, and then saves the
file.
- It re-opens the file, reads, and then verifies the
data.
- It deletes the file.
If any part of the previous operation is unsuccessful, the
Cluster service determines that the hard disk is corrupted and the
ChkDsk /F utility is run to repair possible damage.
The Cluster
service account requires at least
Modify permissions to the
root of the hard disk to perform the previous operations. Entries similar to
the following may appear in the cluster log file (cluster.log) when a
permissions issue causes the Chkdsk utility to run:
Physical Disk Disk Y:DisksWriteTestFile: Creating test file (Y:\zClusterOnlineChk.tmp)
Physical Disk Disk Y:DisksWriteTestFile: CreateFile for (Y:\zClusterOnlineChk.tmp) failed 5
Physical Disk Disk Y:DisksWriteTestFile: Returns status 5
In the previous listing,
Status 5 corresponds to the following error message: