Our setup:
A VMWare VM (5.1) with Windows Server 2008 R2 SP1.
This disk has besides a VMDK also 2 disks with each a direct iSCSI connection to the SAN (Equallogic).
The first iSCSI volume has the production SQL database on it.
The second iSCSI volume has a test SQL database on it.
Sometimes we have to refresh the test database with the production database.
We don't use the standard SQL backup/restore procedure for that, because it would take too long (250GB+ database size).
So we utilize the Equallogic clone volume, by using the Equallogic Snapshot manager:
- We remove/unmount the iSCSI volume with the test database.
- We make a clone of the production volume.
- This clone is mounted again as the test database volume.
Essentially hot-removing/hot-adding a disk.
In this case if you check the Logical Disk performance counters (Free megabytes, %Free space, Disk Queue length etc) for this instance (drive for example F:), you will notice that this data is not available, because the instance isn't shown/available.
The only way so far this can be resolved is by rebooting the server, but that is not always possible. This must be planned.
In the meantime the LogicalDisk counters for this disk aren't recorded by our monitoring system (SCOM 2012 SP1), because this data is not available.
Is this a known issue? Is there a fix?
How to resolve this manually without rebooting the server? So the LogicalDisk performance counters for this disk are available again.
- Edited by Erwin (Eversity) Friday, March 21, 2014 10:23 AM