Since mid August there has been a rebalance issue of the storage in a Windows Server 2019 S2D storagepool. When ever you rebooted a node S2D would kick off a repair moving data off the “down” node. If a node was down for hours this would result in a lot of data being moved off that node.
If you run a 2 node S2D cluster having the correct balance on the nodes is very important. As someone on our storagespacesdirect.slack.com channel experienced yesterday, there 2 Node S2D with Windows Server 2019 was not balanced right and rebooting the wrong node brougt a CSV offline untill the node was back up.
With the Show-Prettypool.ps1 powershell script that you can find here.
In the latest Cumulative Update for Windows Server 2019 KB4476976 this bug is fixed. But as normal, Windows Server bug fixes are seldom listed in the KB articles. But have it confirmed by a trusted source that it is fixed. And as i had this issue and could not rebalance before the 2019 1D patch came out, i can also verify that it works after.
So to manually rebalance run the Optimize-StoragePool -FriendlyName command
One thought on “S2D StoragePool Rebalance bug in Windows Server 2019”
Testing this out with the latest build of 2019 (which is past that update) and still have the same issue. I’m not using servers that support SES but it shouldn’t be required from what I understand. I’m running the optimize storage pool command now and curious if there is a way to see how the data is laid out on the disks to verify one node of a two-node system can actually be shutdown?