srx-chassis-cluster-proxmox

Build and validate Juniper SRX chassis clusters running as Proxmox VE guests.

9|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/fastrevmd-lab/fwskillsshare --skill srx-chassis-cluster-proxmox-fastrevmd-lab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: srx-chassis-cluster-proxmox
Source: https://github.com/fastrevmd-lab/fwskillsshare/tree/main/skills/srx-chassis-cluster-proxmox
Command: npx skills add https://github.com/fastrevmd-lab/fwskillsshare --skill srx-chassis-cluster-proxmox-fastrevmd-lab

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Chassis clusters on Proxmox VE fail silently: the hypervisor substitutes Linux bridges for physical control and fabric links, so clusters form and report healthy while passing no traffic, with nothing logged. This Skill pins the hypervisor-to-Junos seam — bridge and VLAN design, NIC-to-interface mapping, cluster bootstrap, and a validation sequence that catches silent filtering instead of trusting a green status line. ## Core Features & Use Cases - Hypervisor network design: Build a portless, VLAN-aware bridge at MTU 9000 with one VLAN per function, and enforce the load-bearing bridge port flags (learning on, isolation off, neighbour suppression off). - NIC mapping and bootstrap: Map virtual NICs to Junos interfaces (net0→fxp0, net1→em0, shift-by-two for ge- interfaces), configure cluster-id, fab interfaces, reth interfaces, redundancy groups, and static per-node fxp0 addressing under groups. - Failure diagnosis: Symptom-to-cause tables for the anti-spoof trap (per-NIC firewall discarding reth virtual MACs), undersized fabric segments, RG1 hold state, and one-way blackholes from node asymmetry. - Use Case: A two-node vSRX cluster reports all reths Up but passes no traffic. Use this Skill to check the per-NIC firewall setting and the bridge forwarding table for the 00:10:db virtual MACs, identifying the anti-spoof rule as the cause. ## Quick Start Use the srx-chassis-cluster-proxmox skill to plan, build, and validate a two-node vSRX chassis cluster on my Proxmox VE host, starting by checking whether a jumbo-capable bridge exists for the fabric segment.

Frequently Asked Questions about srx-chassis-cluster-proxmox

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a vSRX chassis cluster on Proxmox VE?

Create a portless VLAN-aware bridge at MTU 9000 with one VLAN each for control, fabric, and reth segments, wire identical NIC lists on both guests, then set cluster-id per node and reboot. After boot, configure redundancy groups, fab interfaces, and reth parents, then run the full validation sequence.

Why does my SRX cluster form but pass no traffic?

The most common cause is the per-NIC Proxmox firewall, enabled by default in the web UI. Its anti-spoof rule is pinned to the assigned MAC, so it silently discards every frame using the reth virtual MAC while reths still report Up. Remove firewall=1 from all cluster NICs.

Does the chassis cluster fabric link need jumbo MTU?

Yes, the fabric segment needs MTU 9000 because Junos provisions fab interfaces to 9014/9000 on its own. The control link does not need jumbo. An undersized fabric passes every device health check, so verify the segment MTU on the hypervisor during the build.

Can I enable cluster mode on an existing standalone vSRX guest?

No, a standalone-shaped guest cannot be promoted in place. In cluster mode the second NIC becomes the em0 control link and every remaining interface index shifts by one, breaking all interface references. The NIC plan must be redrawn before clustering.

Why are reth interfaces Down right after committing redundancy groups?

This is the expected RG1 hold-down timer, not a fault. RG1 shows status hold on both nodes with member legs Up and no monitor failures, then elects primary/secondary on its own within a few minutes. Wait for the timer rather than debugging it.

Does this skill cover Juniper Multi-Node High Availability?

No, Multi-Node High Availability is a different HA model with different failure modes and is explicitly out of scope. This skill covers only traditional two-node chassis clusters running as Proxmox VE guests.