S2D Replacing PhysicalDisk Quick Reference

List phsyical disks to find failed disk. Note serial number.
Get-PhysicalDisk
Get-PhysicalDisk | Get-StorageReliabilityCounter

List virtual disks that use the drive, remember them for later.
Get-PhysicalDisk -SerialNumber A1B2C3D4 | Get-VirtualDisk

"Retire" the Physical Disk to mark the drive as inactive, so that no further data will be written to it.
$Disk = Get-PhysicalDisk -SerialNumber A1B2C3D4

$Disk | Set-PhysicalDisk -Usage Retired

S2D should start a to rebuild the virtual disks that utilized the drive.
Get-StorageJob

To be extra safe, run the following on each of the virtual disks that was listed above.
Repair-VirtualDisk -FriendlyName 'VirtualDiskX'

This storage jobs will likely take some time.

Remove the retired drive from the storag pool.
Get-StoragePool *S2D* | Remove-PhysicalDisk –PhysicalDisk $Disk


Physically remove the bad disk.
Physically add a new disk (could peform this first if you have empty drive bays) and check to see if it was added to the storage pool.
Get-PhysicalDisk | ? CanPool –eq True

If nothing is returned it should have been added to the pool, this is what you want as S2D should claim all disks.

If it wasn't added to the pool try the following:
$newDisk = Get-PhysicalDisk | ? CanPool –eq True
Get-StoragePool *S2D* | Add-PhysicalDisk –PhysicalDisks $newDisk –Verbose

Find the new disk's serial number and then see if any virtual disks are using it. None should be yet.
Get-PhysicalDisk -SerialNumber NEWSNBR | Get-VirtualDisk

Rebalance storage pool
Get-StoragePool *S2D* | Optimize-StoragePool
Get-VirtualDisk | Repair-VirtualDisk

Now virtual disks should be using it
Get-PhysicalDisk -SerialNumber NEWSNBR | Get-VirtualDisk

Resize S2D Volume Quick Reference

Get-VirtualDisk vd01
Get-VirtualDisk vd01 | Get-StorageTier | Resize-StorageTier -Size 1.5TB
Get-VirtualDisk vd01


Get-VirtualDisk vd01| Get-Disk | Get-Partition | Get-Volume
$VirtualDisk = Get-VirtualDisk vd01
$Partition = $VirtualDisk | Get-Disk | Get-Partition | Where PartitionNumber -Eq 2
$Partition | Resize-Partition -Size ($Partition | Get-PartitionSupportedSize).SizeMax
Get-VirtualDisk vd01| Get-Disk | Get-Partition | Get-Volume

The Case Against 2-Node S2D Solutions and 2-Way Mirroring

Update: 11/8/2018 Documentation is out for Windows Server 2019 that shows how MS solved the problem. This doesn't solve the problem though if you still decide to utilize 2-way mirroring. Just don't do it. Read on if you want to see what the problem was.


So I've got a two node S2D cluster cooking. The last two times I patched it one of the volumes lost redundancy (just one volume.) The first time it happened I couldn't figure out how to fix it. I ended up blowing the volume away, creating a new one and restoring from backups. This lead me down the path of trying to figure out how to fix this issue in the future which lead to this blog post.

The second time my volume lost redundancy after rebooting a server I thought I was ready for it, since I figure out how to resolve the no redundancy state. Preparing for the worst though, I copied all the VMs off of it so I would have more recent state then from backup. All of the VMs copied except for one. I don't recall the error message it gave me but I think it said something about being unable to read from the disk. This should have been my first clue as to the root cause.

In any case, I had a volume with no redundancy and I attempted the steps I discovered to recover the volume. It didn't work. No matter what I tried. I ended up blowing away the volume again, recreating a new volume and restoring the VMs.

After further investigation it would appear that one of the disks is going bad. I determined this by running the following:

> Get-PhysicalDisk | Get-StorageReliabilityCounter

DeviceId Temperature ReadErrorsUncorrected Wear PowerOnHours
-------- ----------- --------------------- ---- ------------
2                    0                     0    114
5000                                       0    1430
5012                 648                   0    1064
5004                                       0    1417
5006                 0                     0    1051
5010                 0                     0    1064
5009                 0                     0    1051
5003                                       0    1417
5011                 0                     0    1064
5008                 0                     0    1050
5013                 0                     0    1064
5007                 0                     0    1050
5001                                       0    1430

When I run Get-PhysicalDisk all the disks return healthy though. So, the disk is starting to have issues but not enough for the system to think the disk is total garbage yet?

Turns out when I restarted the server without the failing disk, the server WITH the failing disk was the only source of data and it couldn't read form all portions of the failing drive. Hence the no redundancy state. I'm thinking it couldn't read a sector and it couldn't find a redundant copy of the data. Now if this was a three node cluster with 3-way mirroring it could have read from the tertiary copy.

I'm not sure why S2D doesn't take a more proactive approach to resolve the failing disk or at least highlight it more. I'm also not sure why it wouldn't allow me to attach the disk after both nodes were back online. Perhaps the "There is not enough redundancy remaining to repair the virtual disk" warning was because S2D wanted to try and move the data but I needed to add another disk? I was only using 4 TB out of 24TB though, you'd think S2D could move everything off the failing disk to the available space... Perhaps it couldn't attach the disk because changed data could not be replayed from the failing disk to restore the mirror?

I would rather S2D evict the disk right away and make the issue at hand obvious. Or create another health state that indicates a physical disk is in a failing state and trickle that status all the way up to the virtual disks and volumes as unhealthy. Or give us an option to set the number of URE (unrecoverable read errors) threshold for failing a disk.

Long story short, check your disks and make sure none of them are going bad before your reboot servers if you have a two node S2D cluster or if you implement two-way mirroring. Also if you do encounter the no redundancy state it's best to copy as much data off of it as you can before trying to fix it.

Another take away from this is that it would seem wiser to create more smaller volumes instead of fewer larger volumes, of course keeping the number of volumes a multiple of your node count.


Update: I got the following response from Microsoft
"The challenge here is that you had a misbehaving drive… and that’s kind of a gray area.  We handle very well when drives work great… and we handle very well when they fail completely.  But when does bad… become bad enough?  And how do we balance not generating false positives that makes you go replacing drives unnecessarily, and pointlessly wasting money.   With that said, this is an area we are working on.  In Windows Server 2019 we are making enhancements to our Health Service to add what we term marginal drive handling right now (we’ll come up with a better name by ship).

We also hear the feedback that some customers may want higher resiliency out of a 2-node solution, that is another problem we are looking at.  Be mindful that it will come at a cost of reduced efficiency…  but we want to offer customers the choice to do what makes sense for their deployment scenario."

So Microsoft is workign on improving the experience with 2-way mirrors and 2-node S2D deployments. I commend the S2D team, they're very responsive to emails and they listen to what customers have to say. I'm excited to see the improvements with Windows Server 2019.

I just wish there was a way to tweak the algortihm that decides when a drive is bad. I'd personally fail it sooner then later.

Update: Get-PhysicalDisk | Get-StorageReliabilityCounter |  ft DeviceId,ReadErrorsTotal,ReadLatencyMax,WriteErrorsTotal,WriteLatencyMax -AutoSize

Hyper-V SET and NLB in Guest, Duplicate Frames/Packets

Just a quick blurb about switch embedded teaming and Microsoft network load balancing in guest virtual machines. If you've got NLB setup in multicast operational mode running in a VM on top of a Hyper-V host that has SET configured, your VM will receive duplicate frames/packets. It would appear that all the NICs in the SET receive the data and pass it up the network stack. If your upper layer network protocols handle duplicat packets you don't necessarily have to worry about it. ICMP ping for example does not and you will receive multiple responses from your VM. For NLB I would abondon SET and use active/passive NIC teaming.

S2D Recovering a Detached Virtual Disk with No Redundancy

So you've got your S2D cluster and you lost redundancy. Now what? How can this happen? Well if you restart too many nodes this can happen. Lets walk through a sample scenario and see what happens.


Jump to the bottom of the article if you want to skip all the fluff and get right to fix

First the setup:
For this test I setup a 2 node cluster and I created 2 mirrored virtual disks; vd01 and vd02. I setup volumes on them and imported a couple VMs on to each volume. I setup up some VMs so that  the system would be generating IO. Check the Operational Status and the Health Status of the virtual disks by running Get-VirtualDisk. They should all be healthy. I use this script to constantly refresh the virtual disk health and the storage jobs.

Next the failure:
Restart one of the nodes. The virtual disks should have a warning and their operational status should change to a degraded or the degraded,incomplete state. You should also see some repair jobs in the suspended state.

Name   IsBackgroundTask ElapsedTime JobState  PercentComplete BytesProcessed BytesTotal
----   ---------------- ----------- --------  --------------- -------------- ----------
Repair True             00:00:12    Suspended 0               0              29527900160
Repair True             00:04:06    Suspended 0               0              36775657472



FriendlyName ResiliencySettingName OperationalStatus      HealthStatus IsManualAttach Size
------------ --------------------- -----------------      ------------ -------------- ----
vd02                               {Degraded, Incomplete} Warning      True           1 TB
vd01                               {Degraded, Incomplete} Warning      True           1 TB

Once the node comes back online the storage jobs should start running and the disks will be in service. You may see a state of degraded too while storage jobs run. The data that has changed while the node was down is being rebuilt.

Name   IsBackgroundTask ElapsedTime JobState  PercentComplete BytesProcessed BytesTotal
----   ---------------- ----------- --------  --------------- -------------- ----------
Repair True             00:00:04    Suspended 0               0              19058917376
Repair True             00:03:03    Running   16              6276775936     37580963840



FriendlyName ResiliencySettingName OperationalStatus HealthStatus IsManualAttach Size
------------ --------------------- ----------------- ------------ -------------- ----
vd02                               InService         Warning      True           1 TB

vd01                               InService         Warning      True           1 TB

Now before the storage jobs finish rebuilding the redundancy, restart the other node. The disks will likely go into a detached operational status.

Name   IsBackgroundTask ElapsedTime JobState PercentComplete BytesProcessed BytesTotal
----   ---------------- ----------- -------- --------------- -------------- ----------
Repair False            00:00:00    Killed   0



FriendlyName ResiliencySettingName OperationalStatus HealthStatus IsManualAttach Size
------------ --------------------- ----------------- ------------ -------------- ----
vd02                               Detached          Unknown      True           1 TB
vd01                               Detached          Unknown      True           1 TB

You may see a status that says No Redundancy.

Name   IsBackgroundTask ElapsedTime JobState  PercentComplete BytesProcessed BytesTotal
----   ---------------- ----------- --------  --------------- -------------- ----------
Repair True             00:03:12    Suspended 0               0              5368709120



FriendlyName ResiliencySettingName OperationalStatus HealthStatus IsManualAttach Size
------------ --------------------- ----------------- ------------ -------------- ----
vd02                               No Redundancy     Unhealthy    True           1 TB
vd01                               Detached          Unknown      True           1 TB

In this case vd02 is still online but not accessible from C:\ClusterStorage\. If you try to run a repair on the virtual disk with "No Redundancy" you get the following:

PS C:\Users\administrator.SHORELAND_NT> Get-VirtualDisk vd02 | Repair-VirtualDisk

Repair-VirtualDisk : There is not enough redundancy remaining to repair the virtual disk.

Activity ID: {64afdbc9-9ce4-4108-9aac-f4da6d277585}

At line:1 char:24

+ Get-VirtualDisk vd02 | Repair-VirtualDisk

+                        ~~~~~~~~~~~~~~~~~~

    + CategoryInfo          : NotSpecified: (StorageWMI:ROOT/Microsoft/...SFT_VirtualDisk) [Repair-VirtualDisk], CimEx

   ception

    + FullyQualifiedErrorId : StorageWMI 50001,Repair-VirtualDisk

This test system is a clean build with terabytes of storage so it's not a space issue. If your virtual disks only says "No Redundancy" you may want to wait a bit and/or try to offline and then online the disk. This has fixed it before for me.

The same process for recreating the failure will apply for a 3 node or greater setup. Depending on your redundancy though you may have to fail multiple nodes at a time.

For the detached disks, in cluster administrator you will see that the virtual disks are in a failed state. If you try to bring a the virtual disk online through the FCM you'll get an error that says "The system cannot find the drive specified." Error Code 0x8007000F

If you try to connect the virtual disk through powershell you'll get:

Get-VirtualDisk | Where-Object -Filter { $_.OperationalStatus -eq "Detached" } | Connect-VirtualDisk

Connect-VirtualDisk : Access denied

Extended information:
Access is denied.

Recommended Actions:
- Check if you have the necessary privileges to perform the operation.
- Perform the operation from Failover Cluster Manager if the resource is clustered.

Activity ID: {583d3820-dacb-4246-93cf-b52d05d17911}
At line:1 char:82
+ ... -Filter { $_.OperationalStatus -eq "Detached" } | Connect-VirtualDisk
+                                                       ~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : PermissionDenied: (StorageWMI:ROOT/Microsoft/...SFT_VirtualDisk) [Connect-VirtualDisk],
   CimException

    + FullyQualifiedErrorId : StorageWMI 40001,Connect-VirtualDisk

Finally the solution:

Here is what you need to do.

1. Remove all the disks and the pool from the cluster

• In the failover cluster manager select "Pools" on the left under storage
• Select your storage pool in the top pain and then the virtual disks tab on the bottom pain
• Right click each virtual disk and "Remove from Cluster Shared Volumes"
• Right click each virtual disk and "Remove"
• Right click the storage pool in the top pain and "Remove"

2. Go to the Server Manager, File and Storage Services, locate the storage pool, do right click and choose the option “Set Read-Write Access”. Choose one of the nodes. I would choose the server that you have the Server Manager pulled up on. This should allow the single node you selected to have control over the storage pool. This is the key.

3. Select the failed virtual disks, right click and try to attach. It will likely fail but it's going to start the repair process automatically. You can watch the status of the repair with Get-VirtualDisk and Get-StorageJob. Or again you can use the script I created.

Name   IsBackgroundTask ElapsedTime JobState PercentComplete BytesProcessed BytesTotal

----   ---------------- ----------- -------- --------------- -------------- ----------

Repair True             00:00:09    Running  0               0              20937965568

Repair True             00:01:46    Running  16              8020033536     47513075712

FriendlyName ResiliencySettingName OperationalStatus          HealthStatus IsManualAttach Size

------------ --------------------- -----------------          ------------ -------------- ----

vd02                               {No Redundancy, InService} Unhealthy    True           1 TB

vd01                               {No Redundancy, InService} Unhealthy    True           1 TB

The Server Manager should show a warning icon next to the virtual disk while it repairs. You may have to refresh.

4. Once the repair is done the virtual disk's operation status should go to Ok and Health Status to Healthy. The jobs should complete and there should be no more running jobs. The failed and warning icons should go away in the server manager. You may have to refresh. You will be able to attach the virtual disk now. 

5. Re-add the pool and the virtual disks to the cluster again.

• In the failover cluster manager select "Pools" on the left under storage
• Right click "Pools" and select "Add Storage Pool". Select your pool and hit Ok.
• Select your storage pool in the top pain, right click "Add Virtual Disk". Select all of your virtual disk and hit Ok.
• Right click each virtual disk and "Add to Cluster Shared Volumes"

6. Start your VMs back up. You may bave to online the VM's resources if they're not online.

The downside to this process is if you only have one virtual disk that is in a failed/detached/no redundancy state and all the others are fine. You have to take all the virtual disks and the pool out of the cluster to peform the recovery. You may have valid/healthy virtual disks that need to be down (not in the cluster and not being exposed as a CSV) thus you may have VMs that have to be down when their underlying storage is healthy. You could move these to a different location prior to performing the above procedure. Just something to be aware of.

4.20.2018 Updated and Better Solution:
After I wrote this post, I found the following blogpost and gave it a shot. This solution seems to do the trick and it doesn't require taking the storag pool offline or any of your virtual disks that are in perfectly working shape! The gist of it is this:

Remove the disk from CSV. Set diskrunchkdsk and diskrecoveryaction cluster parameters on the disk. Start the disk and start the recovery. Let the recovery job finish. Then stop the disk, revert the cluster paramater settings, add it back to CSV and start the disk.

Remove-Clustersharedvolume -name "Cluster Disk 1"

Get-ClusterResource -Name "Cluster Disk 1" | Set-ClusterParameter -Name diskrunchkdsk -Value 7

Get-ClusterResource -Name "Cluster Disk 1" | Set-ClusterParameter -Name diskrecoveryaction -Value 1

Start-clusterresource -Name "Cluster Disk 1"

Get-ScheduledTask -TaskName "Data Integrity Scan for Crash Recovery" | Start-ScheduledTask

Storage jobs will start, once they finish run:

Stop-clusterresource -Name "Cluster Disk 1"

Get-ClusterResource -Name "Cluster Disk 1" | Set-ClusterParameter -Name diskrecoveryaction -Value 0

Get-Clusterresource -Name "Cluster Disk 1" | set-clusterparameter -name diskrunchkdsk -value 0

Add-clustersharedvolume -Name "Cluster Disk 1"

Start-clusterresource -Name "Cluster Disk 1"

Removing the CSV, adding the CSV and starting/stopping the resource can be done via the FCM

5.11.2018 Update
If you have a 2-node cluster or if you're using 2-way mirroring, read this. You might have a disk that is failing and you might not be able to recover the virtual disk/volume. Check for unrecoverable read errors.

S2D StoragePools and Virtual Disks Can Easily Move

It's possible to move all the physical disks involved in a S2D deployment to different servers and to bring the data online! I tested this out by installing S2D on a three node cluster. I shutdown all the nodes and pulled the disks. I then reinstalled the OS on all three nodes and put the disks back in. I setup the new cluster w/ different server names and cluster name and then ran "Enable-ClusterS2D". The old storage pool and disks could be seen. The storage pool was in a read-only state and the disks were detached. To get the data online I did the following:

Get-StoragePool *s2d* | Set-StoragePool -IsReadOnly $false

Get-VirtualDisk | Connect-VirtualDisk

Then I went to the cluster manager, right clicked on Pools and "Add Storage Pool". Then I right clicked on disks and "Add Disk". At this point I added the disks to CSV and was able to access all the data.

Storage Spaces Direct (S2D), Degraded Virtual Disks and KB4038782

After applying Windows Server 2016 September's patch (KB4038782) to a node in an S2D cluster, the disks on that node would not come out of maintenance mode after the node re-joined the cluster and after Resume -> Fail Roles Back. The VMs would move back but the disks would stay in maintenance mode, thus causing the virtual disks to show a status of degraded. I had to manually take the disks out of maintenance mode after the node joined the cluster and after I failed the roles back. 
To see if the disks are in maintenance mode run:

 Get-StorageFaultDomain

Under the OperationalStatus column it will say "In Maintenance Mode" by the disks for the node that was just restarted. I don't know if this issues was/is just specific to me or if it may happen to everyone that applies the patch. To take the nodes out of maintenance mode use the Disable-StorageMaintenanceMode command.

I have a smaple here that gets all the disks in maintenance mode and disables maintenance mode for those disks.

Update: Official MS KB on the issue is here