NSX Edge upgrade or new deployment fails on 4.1.1 when the upgrade or installation is attempted on older CPUs

Products

• VMware NSX

Issue/Introduction

Symptoms:

• An Edge upgrade or a new deployment of Edges fails on NSX 4.1.1
• Edges datapath fails to come up after an upgrade to NSX 4.1.1
• Datapathd service crashes, and Edges alerts: "Pnic status of the edge transport node <UUID> is down", and "Overall status of the edge transport node <UUID> is down"
• Edge datapath and lb-dispatcher core dump at /var/log/core/

Edge VM or BME CPUs that don't support the AVX2 instruction set are impacted when upgrading to 4.1.1.

• On the Edge, run 'lscpu | grep avx2' to verify if the CPU supports avx2. Example on Edge which does support avx2:

root@edge:~# lscpu | grep avx2
    Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf ljmpq pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single intel_ppin ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm mpx rdt_a avx512f avx512dq rdseed adx smap clflushopt clwb intel_pt avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm arat pln pts pku avx512_vnni md_clear flush_l1d arch_capabilities

 

• In an Edge support bundle, the following command can be run to verify if the CPU supports avx2:

less /proc/cpuinfo | grep avx2

Cause

In NSX 4.1.1 due to the DPDK upgrade on Edges, the machine type for building the DPDK was inadvertently set to 'native' rather than the previously used 'corei7'. This restricted the supported CPU types for the Edge VMs and BMEs to newer CPUs only.

Resolution

This issue is resolved in NSX 4.1.2.

Workaround:
Open a Service Request with Broadcom support to rollback the Edge upgrade.

Additional Information

Impact/Risks:
Edge upgrade or new deployment to NSX 4.1.1 fails.