Performance Summary:

Compliance Errors:

QoS:

•   No Validation Errors Found

Physical Configuration:

Virtual Configuration:

Notes:

• The test sponsor attests, as of date of publication, that CVE-2017-5754 (Meltdown) is mitigated in the system as tested and documented:
• Yes: hosts, guests
• The test sponsor attests, as of date of publication, that CVE-2017-5753 (Spectre variant 1) is mitigated in the system as tested and documented:
• Yes: hosts, guests
• The test sponsor attests, as of date of publication, that CVE-2017-5715 (Spectre variant 2) is mitigated in the system as tested and documented:
• Yes: hosts, guests
• The test sponsor attests, as of date of publication, that CVE-2018-3620 and CVE-2018-3646 (Foreshadow) are mitigated in the system as tested and documented:
• Yes: hosts, guests

Hardware Notes:

Compute Node:

• UEFI settings used:
• Workload Profile: Max Performance Mode, reboot, then choose Custom Mode
• Memory Speed: 2933 MHz (Default 3200 MHz)
• NUMA Nodes Per Socket = 1
• L1 Stream HW Prefetcher = disabled
• L2 Stream HW Prefetcher = disabled
• XCC version: D8BT19M, UEFI version: D8E119J

Storage:

• The Lenovo ThinkSystem DE6000H Hybrid Storage Array provides ESXi datastore for the system under test.
• There are 60 800GB SSD in the DE6000H, all drives are composed as a single pool.
• The pool is powered by Dynamic Disk Pool(DDP) technology, comprising 2 volumes.
• Capacity of volume 1(vol1) is 10TB owned by controller B, and volume 2(vol2) is 20 TB owned by controller A.
• Two volumes are exposed to hosts by FC-NVMe protocol.
• The two FC-NVMe namespaces are assigned to the ESXi cluster that includes the 4 SUT hosts.
• All Departmental VMs and HammerDB VMs were deployed to Vol1, all BigBench VMs were deployed to Vol2

Network:

• Each 100Gb physical switch port is split to 4x25Gb port using 100GbE -> 4 x 25GbE breakout cables

Management Node:

• Each 100Gb physical switch port is split to 4x25Gb port using 100GbE -> 4 x 25GbE breakout cables

Software Notes:

Compute Node:

• Each SUT Host setting for FC-NVMe:
• esxcli system host name set -h ex: esxcli system hostname set --host esxi7-22
• esxcli system module parameters set -m lpfc -p 'lpfc_enable_fc4_type=3'
• Each host was configured with two vswitch, one contains 2 x 25 GbE uplinks for VM traffic and the other contains redundant 2 x 25 GbE uplinks for vMotion traffic
• For redundant vMotion, set two vMotion port group in each host. The override failover order is changed from No to Yes under NIC teaming tab.
• Configure the first adapter is active and the second adapter is standby in the first vMotion port group
• Configure the first adapter is standby and the second adapter is active in the second vMotion port group

Management Node:

• The VCSA is deployed with the configuration that synchronized time with Esxi host and ssh enabled. The deployment size is tiny
• There is 1 datacenter containing two clusters: one for 1 client host and one for 4 SUT hosts
• SUT cluster is configured with DRS enabled, fully automatic
• vSphere DRS Migration Threshold level set to 4
• DRS Advanced Settings: AggressiveCPU Active = 1

Client Driver Notes:

• All svdc clients and the svdc-director were VMs created using SPECvirt Datacenter 2021 template.
• Each svdc client has 4 vCPUs, 12GB memory, 12GB disk, 1 VMware VMXNET Ethernet Controller (rev 01) device and 1 VMware PVSCSI SCSI Controller (rev 02) device
• The svdc-director has 8 vCPUs, 48GB memory, 12GB OS disk, 75GB datastore disk, 2 VMware VMXNET Ethernet Controller (rev 01) devices, and 1 VMware PVSCSI SCSI Controller (rev 02) device
• The second VMware VMXNET Ethernet Controller (rev 01) device on the svdc-director was used as an external management port.
• The client host was configured with two vSwitches:
• vSwitch0 is for management access and uses a 1Gb NIC port
• vSwitch1 is for VM traffic and uses 2 x 25GbE NIC ports

Other Notes:

• None.

Performance Details: