Welcome to (posting on) Communities.
*Generally* if you see some aspect of a host as being detected as a lower version than it is, it is due to lack of proper communication between one of the various daemons on the ESXi side or the services they communicate with in vCenter - as you have this on all clusters it is likely to be on the vCenter side.
Check if vSAN Health (Cluster > Monitor > vSAN > Health) is functional and shows the relevant checks for vSAN iSCSI (e.g. vitd liveness etc.) - if the Health check is blank, not refreshing or not showing full correct information then restart Health and Client services on the vCenter:
# service-control --stop vmware-vsan-health
# service-control --start vmware-vsan-health
# service-control --stop vsphere-client
# service-control --start vsphere-client
Are you 100% positive that all hosts are on 6.7 U1 (build: 10302608) or later and vCenter is indeed on 6.7 U1b as you stated?
Please validate this as 6.7 GA + vSAN 6.7 U1 are not compatible.
If it is only one/few nodes which are failing to retrieve host capabilities and thus triggering this response you may be able to identify this from the vsan health service logs (/var/log/vmware/vsan-health/vmware-vsan-health-service.log), e.g. messages such as this where it shows no capabilities listed::
2019-04-03T11:11:11.11Z INFO vsan-mgmt[Thread-12345] [VsanVcCapabilitySystemImpl::RetrieveHostCapabilities opID=noOpId] Host: host-11, version: 5, capabilities:
The problem have been resolved . I followed your operation tips , restarted the VSAN-health service and vSphere-client ，then iSCSI Taget Services have been recover.
I've seen the log as you said, every host was version: 5 .
Thank you very much.