Failover Cluster - Extremely Slow (CSV Block size?)

[tdh1987]

I’m trying to find out why our failover cluster is so damn slow. I wonder if you would mind telling me what your cluster shared volume block size is and what your experience is like using your VMs?

 

We have 2 x DL380 Gen9 (20 threads, 128GB, 300GB 10K etc), MSA2040 (12Gb SAS connectivity to the hosts), 12 x 1.2TB 10K SAS disks (RAID5) and it's appallingly slow. Trying to install cumulative updates is painful. I'm wondering if my block size is too small.

 

Run these command in command prompt (admin) on one of your hosts:

 

To find your CSV Path:

 

C:\Windows\system32>Wmic volume get caption,blocksize,label

BlockSize Caption Label

4096 C:\

4096 D:\ Quorum

4096 C:\ClusterStorage\Volume1\ CSV01

 

(C:\ClusterStorage\Volume1\ is the path of our CSV)

 

C:\Windows\system32>fsutil fsinfo ntfsinfo C:\ClusterStorage\Volume1\

NTFS Volume Serial Number : 0xac6c3f2c6c3ef126

NTFS Version : 3.1

LFS Version : 2.0

Number Sectors : 0x0000000569538fff

Total Clusters : 0x00000000ad2a71ff

Free Clusters : 0x000000003298bd6e

Total Reserved : 0x0000000000000400

Bytes Per Sector : 512

Bytes Per Physical Sector : 512

Bytes Per Cluster : 4096

Bytes Per FileRecord Segment : 1024

Clusters Per FileRecord Segment : 0

Mft Valid Data Length : 0x00000000000c0000

Mft Start Lcn : 0x00000000000c0000

Mft2 Start Lcn : 0x0000000000000002

Mft Zone Start : 0x00000000000c00c0

Mft Zone End : 0x00000000000cc840

Max Device Trim Extent Count : 1

Max Device Trim Byte Count : 0x40000000

Max Volume Trim Extent Count : 1

Max Volume Trim Byte Count : 0x40000000

Resource Manager Identifier : BEA1558E-B7FE-11E7-A948-E0071B6E23A1

 

Thanks guys

 

Tom

[5tu]

C:\Windows\system32>wmic volume get caption,blocksize,label

BlockSize Caption Label

4096 \\?\Volume{ac11c236-4db9-4b81-8fab-a6be12c948ec}\ Recovery

4096 C:\

4096 Q:\ Quorum

1024 \\?\Volume{9b4a3070-2d4b-4057-8b45-c51d82efb741}\

Z:\

4096 C:\ClusterStorage\Volume2\ Datastore1

4096 C:\ClusterStorage\Volume1\ Datastore2

 

 

C:\Windows\system32>fsutil fsinfo ntfsinfo c:\clusterstorage\volume1\

NTFS Volume Serial Number : 0x6482a8d482a8ac4c

NTFS Version : 3.1

LFS Version : 2.0

Number Sectors : 0x00000001c0028fff

Total Clusters : 0x00000000380051ff

Free Clusters : 0x0000000011df846f

Total Reserved : 0x0000000000000400

Bytes Per Sector : 512

Bytes Per Physical Sector : 4096

Bytes Per Cluster : 4096

Bytes Per FileRecord Segment : 1024

Clusters Per FileRecord Segment : 0

Mft Valid Data Length : 0x0000000000380000

Mft Start Lcn : 0x00000000000c0000

Mft2 Start Lcn : 0x0000000000000002

Mft Zone Start : 0x00000000000c0380

Mft Zone End : 0x00000000000cc840

Max Device Trim Extent Count : 0

Max Device Trim Byte Count : 0x0

Max Volume Trim Extent Count : 62

Max Volume Trim Byte Count : 0x40000000

Resource Manager Identifier : E9310D43-185E-11EB-AD92-FEEBDAB97D00

[timbo343]

Hi Tom

 

We have kinda the same setup too but a little bit older yet experiencing the same symptoms - 2x HP Dl380 G8 (128GB RAM in each, mirrored 300Gb 10K drives) running server 2012R2 - our SAN is a DotHill which i've been told this SAN is the same chassis as the HP MSAs. The only thing is we have 12x 1tb 7200k disks in a RAID5 in ours. Our SAN has no errors or issues to report.

 

Here is out block level report on our Cluster Volume

 

Edited January 21, 2022 by timbo343

[5tu]

Same block size and our cluster is nippy so don't think that's your issue.

[Steve21]

On your info your physical sector is only 512 on those HDDs rather than the cluster size

 

But don’t forget it could also be things like your raid controller cache, what are you running on that side?

 

Steve

[tdh1987]

On your info your physical sector is only 512 on those HDDs rather than the cluster size

 

But don’t forget it could also be things like your raid controller cache, what are you running on that side?

 

Steve

 

Hi Steve

 

Are you referring to the cache size on the controller (MSA2040)? If so, that's 6144MB

 

Thanks

 

Tom

[Oaktech]

How are you connecting to it? Is it possibly a SAN network card driver issue?

[tdh1987]

How are you connecting to it? Is it possibly a SAN network card driver issue?

 

Hi

 

We're using the H241 SAS controller. Since posting, I've updated the firmware from 5.52 to 7.00 and also updated the driver. Let's hope it improves performance