Hello Grim192
"I decided to upgrade to v6.7. The vCenter upgraded fine but the hosts kept failing, so I decided to just wipe the hosts and the storage and start fresh. Each of the hosts were freshly installed with v6.7"
Please confirm the hosts and vCenter are on the same version e.g. both on 6.7 GA/pre-U1 or both on 6.7 U1 - I say this as there have been a LOT of backend changes made in 6.7 U1 and as a result have seen some loopy behaviour when trying to create/administer 6.7 U1 vSAN clusters with a 6.7 vCenter (they are incompatible and unsupported).
"If I rebooted host 1 the boot would hang until I removed the SSD and then it booted successfully"
Where did this get stuck? In the module pre-load or during after this (and at what module)?
"2018-12-30T00:03:55.487Z cpu0:2097175)WARNING: NMP: nmpCompleteRetryForPath:357: Retry cmd 0x28 (0x459a40d2c140) to dev "t10.ATA_____V42DCT064V4SSD2__________________________200105984___________" failed on path "vmhba1:C0:T0:L0" H:0x1 D:0x2 P:0x$"
Looks like you maye be using 'less' and cutting off the end of lines, try using 'cless' or less with line-wrap switch.
Please recreate the issue via GUI, retry using CLI (e.g. esxcli vsan storage add -s naa.xxxxxx -d naa.xxxxxxx) and then pull the vobd.log and vmkernel.log and attach them here.
You could also try switching the devices into one of the other hosts and testing if the problem follows it (indicating a hardware issue).
Bob