vmx crash with "mks - PANIC: Unexpected signal: 11." when connecting multiple consoles

Products

• VMware vSphere ESXi

Issue/Introduction

• In rare circumstances, connecting multiple VM consoles while the VM is changing it's screen resolution can result in a vmx crash.
  
  
  
• Event logs show entries similar to
  
  An application (/bin/vmx) running on ESXi host has crashed (1 time(s) so far).
vSphere HA initiated a failover action on <VM> virtual machines in cluster <Cluster> in datacenter <Datacenter>
  
  
  
• The vmware.log shows that the resize operation of the existing console and the opening of a second console occurred almost simultaneously, and a signal 11 was received.
  
  [timestamp] In(05) svga - MVNCEncode: MVNCEncode_ServerSetTopology - original root: (0, 0) size: (1024, 768)
 ::
[timestamp] In(05) vmx - VigorTransportProcessClientPayload: opID=<opID> seq=79293: Receiving MKS.IssueTicket request.
 ::
[timestamp] In(05) mks - MKSRemoteMgr: AddRemoteConsole (numConsoles=1)
[timestamp] Wa(03) mks - Caught signal 11 -- tid ########## (addr C)
 ::
[timestamp] Cr(01) mks - PANIC: Unexpected signal: 11.
[timestamp] In(05) mks - Panic: can't get userlevel lock.
[timestamp] Wa(03) mks - A core file is available in "/vmfs/volumes/path/to/vmx-zdump/vmx-zdump.###"
  
  
  

Cause

A race condition between resizing the existing console and attaching a second console can occasionally trigger this.

Resolution

A fix is planned in ESXi 8.0.3 P09 release.

As a workaround, avoid opening multiple consoles.

If you want to limit the number of consoles, you can set the RemoteDisplay.maxConnections option.
For configuration details, see the "RemoteDisplay.maxConnections" section in the following document:

Virtual Machine Security Controls Reference