Claimrules added using 'vendor', 'model' option are not claiming the NVMe devices as expected.
Example: Following claimrule for vendor model 'Vendor-name*', might not claim NVMe devices.
esxcli storage core claimrule add --rule 123 -t vendor -P HPP -V NVMe -M “Vendor-name*” --config-string “pss=LB-Latency,latency-eval-time=120000"
NVMe devices present vendor and model information in identify controller response as a single field and this cannot be split as separate fields. With SCSI<->NVMe translation stack we could support this based on the standard inquiry response based on SCSI<->NVMe translation reference specification, but now ESXi supports native NVMe stack and does not follow the SCSI<->NVMe translation model. Hence, VENDOR and MODEL cannot be supported as separate fields.
For NVMe devices, use the "nvme_controller_model" parameter when using vendor and model based claimrules.
Instead of nvme-controller-model 'Vendor-name*', use --nvme-controller-model 'CONTROLLER MODEL NAME <regular expressions allowed>' for NVMe device claimrules.
Example: esxcli storage core claimrule add --rule 123 -t vendor --nvme-controller-model 'Vendor-name*' --config-string “pss=LB-Latency,latency-eval-time=120000"