2 node vsan robo partioned and cannot use one host

[bigslash]

hi to all

we have 2node branch office VSAN solution ,our witness server is in a central location.we have some alerts and prolems,When we check the Health of the Vsan ,we see the 2 node are in partion 1,and vsan witness is in partion2 .Is this normal? 

I had checked the network with vmkping command between nodes and witness ,i see i can reach all of them .But still got the error partioned.And because of that error i guess, the Node is not replicating the vmdk files with other Node (raid 1)

ı had checked this below,(seems similar but not same) 

"https://kb.vmware.com/s/article/89087"

This is production site, so i need to be very carefull, what i had done till is below

1-Checked with vmkping all traffic between nodes and witness server

2-restarted all agents in non replicating node

3-rebooted the none replicating node

4-Rebooted the witness server

none helped yet, 

thanks

[drdoc_fred]

If you are unable to use one host, it might be due to the version of vCenter Server and the witness host. For eg, vCenter Server must be version 8.0 Update 2 later requires the witness host must be ESXi version 7.0 Update 2 and later.

[BarryGrowler · Accepted solution]

To resolve the partitioning issue in your 2-node VSAN ROBO setup, start by verifying network connectivity between nodes and the witness, and check the VSAN health check results for any specific errors.

Ensure the witness components are running properly and resynchronize VSAN objects using the "Repair Objects" feature. Also I recommend to restart VSAN services on the affected nodes and check the health of the physical disks.

[bigslash]

we solved our issue with deploying new witness servers to local site

we cannot pass the firewalls between the ROBO vsan and the virtual center site (the witness server was here)

to test the issue , we used this KB (Troubleshooting vSAN Witness Node Isolation (vmware.com))

mkping -I <VSANvmknic> <WitnessIP> -s 1472 -d -c20

Thanks BArry 🙂