vSAN network mode

[suhag79]

Hi,

We have recently configured the fresh vSAN cluster. Below are the summary. wanted to know, will that be any issue with Networking Mode as Unicast ?  Though overall state is healthy.

[daphnissov]

Not if health checks are returning successful. In vSAN 6.6, the mode was automatically changed to unicast from multicast, and there's nothing you need to do afterwards.

[TheBobkin]

Hello suhag79

If all nodes are on 6.6 build of ESXi/vSAN and have disks on latest on-disk format as they do appear to have in your screenshot then Unicast as the mode of communication is expected.

You can easily verify the network communication and connectivity between vSAN nodes in a number of ways:

From the Health UI in the vSphere Web Client:

Cluster > Monitor > Health > Network drop-down  - if these are green for the main tests then network is likely fine.

From the CLI via SSH:

Check the Cluster membership includes all nodes:

# esxcli vsan cluster get

Check that their Unicast lists are fully populated with all other nodes in the clusters info:

# esxcli vsan unicastagent list

Check ping between the vmk configured on one host to the others e.g.:

# vmkping -I vmkX 192.168.X.X

(get the IP and vmk# of the vSAN interface from GUI or #esxcfg-vmknic -l)

You can also check that creation of VMs/Objects that are FTT=1 (Default Storage Policy) is possible by manually creating a VM or using the VM creation test under proactive tests - this would not be possible without network connectivity between enough hosts to satisfy the placement requirements of the SP (3 nodes/FDs).

Bob

[suhag79]

Thanks TheBobkin,

network is green but i need to work on other three warning.

We have updated the drive but still it is showing Hardware compatibility warning.

[TheBobkin]

Hello suhag79​,

Latest drivers available are often not the vSAN-certified drivers (and firmware) so you need to check against the vSAN specific HCL to verify you are using drivers compatible with vSAN.

Identify the controller (e.g vmhba1) and driver name (e.g. lsi_mr3):

# esxcfg-scsidevs -a

Check the driver loaded:

# vmkload_mod -s <driver name> |grep Version

Check the Vendor ID (VID), Device ID (DID), Sub-Vendor ID (SVID), and Sub-Device ID (SDID) of the controller to verify the exact model:

# vmkchdev -l |grep vmhbaX

Enter these IDs into the vSAN HCL for controllers or use this tool my colleague created:

https://hcl.captain-vsan.com/

This will tell you what drivers and firmware you should be using on the controller.

What are the specifics showing as warnings in the drop-down menus for 'Cluster' and 'vSAN Build recommendation' and 'Online Health' showing as warning?

Bob

[suhag79]

Thanks, will try that

Below are the warning for cluster

below are the warning for vSAN Build recommendation. i think, i can ignore this one ?

Online Health

[TheBobkin]

Hello suhag79​,

'Time is synchronized across hosts and cluster':

- Set up NTP correctly across the cluster (vCenter and hosts)

https://kb.vmware.com/s/article/2012069

Information about this check:

https://kb.vmware.com/s/article/2149505

vSAN Build Recommendation Engine health:

- This can have multiple causes but most likely are: VUM, internet connectivity and logged into my.vmware.com

Information about this check:

https://kb.vmware.com/s/article/2150914

CEIP

- either join CEIP (if you want) or disable this health check:

Information about this check:

https://kb.vmware.com/s/article/2148866

How to disable Health checks:

http://www.virten.net/2017/04/how-to-silence-vsan-health-checks/

Bob

[suhag79]

Regarding controller drive version, i have checked and as per vSAN HCL, they are advising to back to much lower version. Is it ok to remain at higher level of driver or will it create any issue ?

[TheBobkin]

Hello suhag79​,

It is extremely important that you only use driver/firmware for controller that are certified for the version of vSAN in use - using unsupported drivers that are NOT on the vSAN-HCL for these controllers is a very bad idea and could potentially result in a number of issues ranging in severity from mild to catastrophic.

Downgrade these drivers to the current latest driver certified for this controller as per the vSAN-HCL (and firmware too if the current in use is not compatible with the driver being downgraded to).

Bob

[suhag79]

Agreed.