VM with Fault Tolerance enabled terminates Fault Tolerance protection spontaneously

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:
Fault Tolerance fails over with error messages in vmkernel log similar to:

vmkernel.2:2018-05-05T15:45:10.647Z cpu31:9696746)WARNING: FTCpt: 712: (4893094829365 snd) Too many dirty pages: allowed 262144, msg 469676: Limit exceeded...

Note: The preceding log excerpts are only examples. Date, time, and environmental variables may vary depending on your environment.

Environment

VMware vSphere ESXi 6.0

Resolution

To resolve this issue, increase the value of ftcpt.maxDirtyPages from the default value of 262144 (1GB memory). 

1. Observe the value indicated in the msg XXX portion of the log message.

Example: msg 469676: Limit exceeded

2. Add 125000 or 245000 to this value (corresponding to adding 512MB or 1GB to the limit).

Note: Increasing the value will proportionately increase the memory usage for the VM.

3. Set this new value to the ftCpt.maxDirtyPages parameter. For more information on how to add a new configuration parameter, see Modifying advanced virtual machine settings using vSphere Client.
4. Power off/power on the VM (or vMotion the VM) to allow the changes to take effect.

Note: If the issue persists, repeat steps 1 - 4.