"upstream connect error or disconnect/reset before headers" error when accessing SRM/vSphere Replication/dr page URL

Products

• VMware Live Recovery

Issue/Introduction

Symptoms:

• Site Recovery Manager (SRM)  or vSphere Replication  (VR) plugin in vCenter ( VC) reads  "UI error: Unknown error"
  
  
• On accessing the /dr URL page manually we receive browser error, ie: https://Appliance_FQDN/dr 
  
  
• Error seen - "upstream connect error or disconnect/reset before headers. reset reason: connection failure, transport failure reason: delayed connect error: 111"
  
   

Environment

VMware vSphere Replication 

VMware Live Site Recovery

Cause

• The appliance dr-client service is hung or in stopped state. 
  
  Example:
  
  
  Note: Since the Startup Type of the service is "Manual", it will not auto start the service post appliance reboot. So, this service needs to be started manually. dr-client service is responsible from SRM/DR UI on the browser.
  
  

  An RFC 1123 hostname compliance violation where the VLR appliance FQDN contains UPPERCASE characters causes initialization and routing failures within the internal Envoy proxy network fabric (no healthy upstream). When the HTTP engine cannot contact a live listening socket for the replication daemon, it defaults to an explicit 503 Service Unavailable response.

  Additionally, the integrated Envoy proxy services inside VLR 9.x strictly enforce lowercase string matching for structural endpoint routing; UPPERCASE symbols disrupt the lookup mechanism, resulting in an immediate breakdown of the ingress path and triggering a no healthy upstream error.  

Resolution

Follow below steps -

• Launch Site Recovery Manager (SRM)  or vSphere Replication VR) Manager VAMI 5480 page
  
  
• Navigate to Services Tab
  
  
• Restart the dr-client service
  
  
  
  Check and modify the appliance hostname FQDN to lowercase to resolve the casing evaluation rules within the Envoy web server.