When Lustre servers enable 'suppress_pings', all clients will stop pinging. The servers will rely on external node failure notification mechanism to evict dead client through the procfs evict_client. However, for Lustre servers serving multiple client systems, client systems may have mixed capability of delivering the node failure event. For those clients that may not have such mechanism available need to preserve the Lustre ping. One way to resolve the issue is to add a client mount option 'always_ping' so that the client will not stop pinging even server enabled 'suppress_pings'.