This article provides information on how to troubleshoot a clustering issue, when running over a L2 switch. It also is a jumpstation to other useful KB links on this topic.
How to troubleshoot a clustering issue, when running over a L2 switch
First, when clustering over a L2 switch, you do not have to implement any extra configuration/settings on the SRX device; as compared to clustering with a back-to-back connection. If removing the switch resolves the issue, then the issue is with the switch. Procedure on the switch : As the SRX is using a Juniper proprietary protocol for HA communication, confirm the following parameters:
TN21 - SRX High Availability over Layer 2 Infrastructure Guide (login required) KB23995 - [SRX] How to form a SRX240 cluster over EX2200 switches [juniper.net] KB24843 - How to enable or disable VLAN tagging on chassis cluster control port? [juniper.net] KB15141 - Secondary SRX node is always going to disabled state when fabric link connection is done via Cisco 4500/6500 switch [juniper.net] KB20698 - [SRX] Troubleshooting steps to correct a Control Link down situation in a Chassis Cluster [juniper.net] KB20687 - [SRX] Troubleshooting steps to correct a Fabric Link that is down in a Chassis Cluster [juniper.net]
2020-07-24: Fixed broken link. 2020-03-24: minor non-technical edits.