J
Jim Riley_GBC
Hello,
We recently purchased a new server which we will be using to support several of our SQL databases. The server contains 4 NVMe drives that will be used for the database files. The server is running Windows Sever 2019. We would like to use Storage Spaces to “manage” these drives. We didn’t have any experience with Storage Spaces, so we have been doing some testing to make sure we understand how to setup and manage it.
Everything is working as expected, except for replacing failed drives. In order to test replacing a failed drive, we did the following. First we created a storage pool using two of our NVMe drives. We then created a mirrored virtual disk which used all the space in that pool. We created some test databases on this virtual drive and ran a couple of quick tests. Using Performance Monitor, we could see that Storage Spaces was balancing the IO between the two drives in the pool.
We then pulled out one of the drives to simulate a drive failure. Since the virtual drive was mirrored, everything worked fine and the databases had no issues. We then “retired” the bad drive and used the Add-PhysicalDisk command to add one of our other NVMe drives into the pool. After the Storage Spaces job successfully completed, we then used the Remove-PhysicalDisk to get rid of the bad disk.
Both the storage pool and the virtual disk reported a status of Healthy\OK. We then ran a couple more database tests and while we didn’t get any errors, we did notice that in Performance Monitor the newly added drive was never being read from during our tests. It was being written to at the same rate as the other drive, but all of the read IO was being handled by the other drive in the storage pool (i.e. Storage Spaces was not balancing the read IO as it did when we first created the storage pool\virtual drive).
We tried a couple of things to see if we could get Storage Spaces to balance the read IO again. That included: rebooting the server, running the Optimize_StoragePool command and running the Repair_VirtualDisk command. None of that fixed the issue. Does anyone know if this is normal behavior and are there any commands that we can use to get Storage Spaces to rebalance the read IO as it did when we initially created the storage pool\virtual disk?
Any insight\advice would be greatly appreciated.
Thank You
Continue reading...
We recently purchased a new server which we will be using to support several of our SQL databases. The server contains 4 NVMe drives that will be used for the database files. The server is running Windows Sever 2019. We would like to use Storage Spaces to “manage” these drives. We didn’t have any experience with Storage Spaces, so we have been doing some testing to make sure we understand how to setup and manage it.
Everything is working as expected, except for replacing failed drives. In order to test replacing a failed drive, we did the following. First we created a storage pool using two of our NVMe drives. We then created a mirrored virtual disk which used all the space in that pool. We created some test databases on this virtual drive and ran a couple of quick tests. Using Performance Monitor, we could see that Storage Spaces was balancing the IO between the two drives in the pool.
We then pulled out one of the drives to simulate a drive failure. Since the virtual drive was mirrored, everything worked fine and the databases had no issues. We then “retired” the bad drive and used the Add-PhysicalDisk command to add one of our other NVMe drives into the pool. After the Storage Spaces job successfully completed, we then used the Remove-PhysicalDisk to get rid of the bad disk.
Both the storage pool and the virtual disk reported a status of Healthy\OK. We then ran a couple more database tests and while we didn’t get any errors, we did notice that in Performance Monitor the newly added drive was never being read from during our tests. It was being written to at the same rate as the other drive, but all of the read IO was being handled by the other drive in the storage pool (i.e. Storage Spaces was not balancing the read IO as it did when we first created the storage pool\virtual drive).
We tried a couple of things to see if we could get Storage Spaces to balance the read IO again. That included: rebooting the server, running the Optimize_StoragePool command and running the Repair_VirtualDisk command. None of that fixed the issue. Does anyone know if this is normal behavior and are there any commands that we can use to get Storage Spaces to rebalance the read IO as it did when we initially created the storage pool\virtual disk?
Any insight\advice would be greatly appreciated.
Thank You
Continue reading...