Server 2016 with two nodes + failover cluster manager

vdes2020

New Member
Jul 12, 2020
12
0
1
Hi All,



Hoping someone can assist. We've got two physical servers running 2016 running running a failover cluster manager environment. Each server has the following hard disks installed.

8 x 600GB 15k SAS on storage controller 1

2 x 600GB 15k SAS (mirrored as used by host OS) on storage controller 2

4 x 960GB SSDs on storage controller 2

We've got one storage pool presented to failover cluster manager and then running two CSV's. One of the CSVs was running out of space therefore we decided to install an additional 2 x 960SSDs on each server to provision more space.

When we check the storage pool I can see it says 2.18TB of free space. When I go to extend the CSV's with the assumption I've got 2.18TB of free space, I cannot increase the size of the CSV by that amount - at the moment the only amount I can extend it by is 50GB which is not making a lot of sense to me. Why? I've already optimized the storage pool with little effect. I've also rescanned the the storage pool again with little effect.

Please see below some outputs:

PS C:\> get-storagesubsystem clu* | get-storagehealthreport
CPUUsageAverage : 8.86 %
CapacityPhysicalPooledAvailable : 2.18 TB
CapacityPhysicalPooledTotal : 20.3 TB
CapacityPhysicalTotal : 20.3 TB
CapacityPhysicalUnpooled : 0 B
CapacityVolumesAvailable : 556.78 GB
CapacityVolumesTotal : 4.84 TB
IOLatencyAverage : 185.59 us
IOLatencyRead : 142.95 us
IOLatencyWrite : 1.29 ms
IOPSRead : 14455.62 /S
IOPSTotal : 15014.48 /S
IOPSWrite : 558.86 /S
IOThroughputRead : 1.29 MB/S
IOThroughputTotal : 9.31 MB/S
IOThroughputWrite : 8.02 MB/S
MemoryAvailable : 109.15 GB
MemoryTotal : 384 GB




PS C:\> get-physicaldisk

FriendlyName SerialNumber CanPool OperationalStatus HealthStatus Usage Size
------------ ------------ ------- ----------------- ------------ ----- ----
HP EH0600JDYTL 0XKAN7LP False OK Healthy Auto-Select 558.91 GB
ATA VK000960GWTHB SJ95N7459I0205T5G False OK Healthy Auto-Select 894.25 GB
HP EH0600JDYTL 0XKAU4WP False OK Healthy Auto-Select 558.75 GB
ATA MK0960GECQK S1JANWAG604139 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKARTBP False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKARPTP False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAPW3P False OK Healthy Auto-Select 558.75 GB
ATA MK0960GECQK S1JANYAG703607 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKAU93P False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XK86KTR False OK Healthy Auto-Select 558.75 GB
ATA MK0960GECQK S1JANYAG703371 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XK86KPR False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XK86JMR False OK Healthy Auto-Select 558.75 GB
ATA MK0960GECQK S1JANYAG703392 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKAU0NP False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAL58P False OK Healthy Auto-Select 558.75 GB
ATA VK000960GWTHB SJ95N7459I0205T42 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKAU4BP False OK Healthy Auto-Select 558.75 GB
ATA VK000960GWTHB SS95N4482I0105T60 False OK Healthy Auto-Select 894.25 GB
ATA MK0960GECQK S1JANYAG703466 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKAN6UP False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XK86K5R False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAV1XP False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAD0WP False OK Healthy Auto-Select 558.75 GB
ATA VK000960GWTHB SJ95N7459I0205T0U False OK Healthy Journal 894.25 GB
ATA MK0960GECQK S1JANYAG703401 False OK Healthy Journal 894.25 GB
ATA MK0960GECQK S1JANYAG703326 False OK Healthy Journal 894.25 GB
ATA MK0960GECQK S1JANYAG703421 False OK Healthy Journal 894.25 GB
HP EH0600JDYTL 0XKA663P False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAR03P False OK Healthy Auto-Select 558.75 GB
HP EH0600JDYTL 0XKAV1LP False OK Healthy Auto-Select 558.75 GB



PS C:\> Get-StoragePool

FriendlyName OperationalStatus HealthStatus IsPrimordial IsReadOnly
------------ ----------------- ------------ ------------ ----------
Primordial OK Healthy True False
Primordial OK Healthy True False
S2D on print-cluster1 OK Healthy False False



PS C:\> .\show-prettypool.ps1

Volume Filesystem Capacity Used Resiliency Size (Mirror) Size (Parity) Footprint Efficiency
------ ---------- -------- ---- ---------- ------------- ------------- --------- ----------
CSV-01 ReFS 2TB 96% 2-Way Mirror 2TB 0 5TB 50%
CSV-02 ReFS 2TB 84% 2-Way Mirror 2TB 0 5TB 50%


Can someone please explain and recommend a solution here as I need to be able to used those new SSDs to increase the space on the CSV.

In the output above, the disks labelled "ATA VK000960GWTHB" are the new ones.
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
What is your column count of your virtual disks?

Get-VirtualDisk | ft FriendlyName, ResiliencySettingName, NumberOfColumns, NumberOfDataCopies, @{Expression={$_.Size / 1GB}; Label="Size(GB)"}, @{Expression={$_.FootprintOnPool / 1GB}; Label="PoolFootprint(GB)"} -AutoSize

I am assuming that this is an S2D cluster?

Chris
 

vdes2020

New Member
Jul 12, 2020
12
0
1
Hi,

Thanks for your response. Please see below:


1594765642594.png

Yes correct this is part of a S2D cluster.
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
Here is the bad news... because the column count is 8 for both volumes, you have to add 8 disks to the (each) pool to enlarge the VDisk (CSV)...

Chris
 

vdes2020

New Member
Jul 12, 2020
12
0
1
Oh, no :(

So possibly stupid question but what did it achieve when I added 2 x 1TB SSDs on each server?

Also, when the storage pool shows me 2.18TB of free space, what exactly is that? i.e. what's the purpose of showing me that if I cannot use it?
 
Last edited:

vdes2020

New Member
Jul 12, 2020
12
0
1
Hi - yep I ran that a few days ago but it made no difference unfortunately.
I can run it again if you like? But I ran it twice already.
 

vdes2020

New Member
Jul 12, 2020
12
0
1
Unfortunately that's not going to be possible as the server has no more slots :(

Do you know what did it achieve when I added 2 x 1TB SSDs on each server? Seems adding them or not made no difference ?
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
You are past the point of safety. You do not have any wiggle room to do anything. I would migrate vms off to other servers then rebuild the csvs with the additional and/or get larger disks.

That is a temp solution. For now.

Chris
 

vdes2020

New Member
Jul 12, 2020
12
0
1
Ok thanks.

is it possible to reduce the size of CSVs? I’m thinking of moving non-critical VMs off the clusters then reduce CSV01 and assign CSV02 more space as that needs it more.
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
you cannot reduce the size of CSV's as they are fixed allocation. and the 2 TB free is referring to the 2 disks you added... but the volumes cannot expand into them as you do not have enough disks for the VDisk column count.

optimize-storagepool is supposed to help here. but apparently it is not.

Chris
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
Migrate all of the ownership of the pool and the CSVs to one node and then run optimize-storagepool on that node. let's see if that helps.

Chris
 

vdes2020

New Member
Jul 12, 2020
12
0
1
OK so I can see the owner of the pool is the second host.
CSV01 is host 1
CSV02 is host 2

I'm assuming I can make host 2 owner of CSV01 without downtime/outage is that right?
 

cesmith9999

Well-Known Member
Mar 26, 2013
1,202
362
83
Correct there is no downtime in transferring ownership. pool and CSV ownership is handled in 2 different places in the failover manager just right click and transfer ownership.

Ownership in Hyper-V failover cluster is like being a traffic cop... the owner is the one that allows/pauses writes from different VM's on different nodes to be done in a sequential manner. reads are generally always allowed.

Chris
 

vdes2020

New Member
Jul 12, 2020
12
0
1
Confirming I've moved the pool + CSVs to host 2, run the optimize command on the pool on host 2, no difference made.
 

vdes2020

New Member
Jul 12, 2020
12
0
1
If I take out those 4 new SSDs and replace them with normal 15K SAS drives (i.e. the same ones as what's currently in them), will that allow me to use the new disks as capacity? or will that not be possibly because of the columns?
 
Last edited: