With the latest TrueNAS (13U6) and ESX 7.8 and 8.0, is NFS still a easily solution than ISCSI for VM? This is for a non-production environment where most of the VM provisioning, installation, testing, etc are either automated or no critical data would be kept. In short, performance prioritize over data protection. The TrueNAS HW spec listed below
AMD Ryzen Threadripper 1920X 12Core 3.5GHz
Gigabit MB with 1 LOM GB NIC
128GB of RAM (Max out)
LSI SAS 9300-16I 12GB/S HBA (in IT Mode) to 8 Seagate EXOS 8 TB HD
LSI SAS 9300-8i 12GB/S HBA (in IT mode) to Samsung 870 EVO 4 TB (2 now, plan to scale up to 8 total)
Boot drive - 128GB SSD 3D TLC NAND PCIe Gen 3 x2 NVMe M.2
2 - Intel® Optane™ SSD P1600X (58GB) plan for SLOG or L2ARC for the HDD pool (not in used yet)
Dual port Mellanox CX314A ConnectX-3 Pro
Dell PowerConnect 8132F Switch
Intel x540 Dual 10GB on ESX host, running to the same Dell PowerConnect 8132F Switch.
The plan is to allocate all the SSD for VMware workload (datastore for VM, not production critical), and all the HDD for production critical data (backup files, configuration data, scripts, video, etc). For starter, given the server maximum memory top out at 128GB, would it still be better to go with NFS versus iSCSI ?
Second, can I just create a pool with all 8 ssd (each as its own vdev) and strip them together to get the maximum performance?
Third, given it is SSD, would L2ARC or SLOG matter here?
Any input would be much appreciated
AMD Ryzen Threadripper 1920X 12Core 3.5GHz
Gigabit MB with 1 LOM GB NIC
128GB of RAM (Max out)
LSI SAS 9300-16I 12GB/S HBA (in IT Mode) to 8 Seagate EXOS 8 TB HD
LSI SAS 9300-8i 12GB/S HBA (in IT mode) to Samsung 870 EVO 4 TB (2 now, plan to scale up to 8 total)
Boot drive - 128GB SSD 3D TLC NAND PCIe Gen 3 x2 NVMe M.2
2 - Intel® Optane™ SSD P1600X (58GB) plan for SLOG or L2ARC for the HDD pool (not in used yet)
Dual port Mellanox CX314A ConnectX-3 Pro
Dell PowerConnect 8132F Switch
Intel x540 Dual 10GB on ESX host, running to the same Dell PowerConnect 8132F Switch.
The plan is to allocate all the SSD for VMware workload (datastore for VM, not production critical), and all the HDD for production critical data (backup files, configuration data, scripts, video, etc). For starter, given the server maximum memory top out at 128GB, would it still be better to go with NFS versus iSCSI ?
Second, can I just create a pool with all 8 ssd (each as its own vdev) and strip them together to get the maximum performance?
Third, given it is SSD, would L2ARC or SLOG matter here?
Any input would be much appreciated