ESXi PSOD with #PF Exception 14 in Lwd_GetDemandLogInfo (hbr_filter)

Products

• VMware vSphere ESXi

Issue/Introduction

• ESXi PSOD with #PF Exception 14 with Module(s) involved in panic: hbr_filter
• PSOD Message may be similar to below: 

         #PF Exception 14 in world #######: PreemtibleT IP ######### addr ####

• ESXi PSOD experienced with back trace similar to below:

  #0  ################## in Lwd_GetDemandLogInfo (diskInfo=diskInfo@entry=##################) at .....
  #1  ################## in DemandLogRecycleIO (io=io@entry=##################) at .....
  #3  ################## in DemandLogTriggerAbortAndFinalizeExtent (data=##################, timestamp=<optimized out>) at .....
  #4  ################## in TimerWheelHandlerLoop (list=##################, curTC=##################, t=##################) at .....
  #5  ################## (t=t@entry=##################, endTC=endTC@entry=##################) at .....
  #6  ################## in TimerWorld (clientData=<optimized out>) at .....
  #7  ################## in CpuSched_StartWorld (destWorld=<optimized out>, previous=<optimized out>) at .....
  #8  ################## in ?? () at .....
  #9  ################## in ?? () 

 

 

 

Environment

VMware vSphere ESXi 8.0 U3g and below 

Cause

This issue is caused by a software race condition within the replication filter (hbr_filter) module. 

The race condition is because of the filter detach and disk info callback overlaps for the same disk from 2 different functions.

 

Resolution

Our engineering team is aware of this issue and are working on a permanent fix 

Workaround: 

• Disable quiescing for all replication VMs to try and avoid the frequent hbr filter attach/detach which lead to PSOD 
• This workaround will reduce the chance of hitting this race condition until a fix is released 

 

Subscribe to this article to be notified of latest updates