CIFS FlexGroups with ChangeNotify become unresponsive

Applies to

• ONTAP 9.2 - 9.7P15
• CIFS Change Notify
• SAN/NAS protocols

Issue

• Multiple seconds of latency is seen on FlexGroups
• Node communication becomes unresponsive impacting SAN and NAS protocols
• The virtual machine (VM) or application on the LUN may be down and unable to connect to the LUN
  • High latency is seen like with qos statistics:

    Cluster::> qos statistics latency show -node node1
    Policy Group            Latency    Network    Cluster       Data       Disk    QoS Max    QoS Min      NVRAM      Cloud  FlexCache    SM Sync         VA
    -------------------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ----------
    -total-                  10.44s     10.44s    38.00us        0ms        0ms        0ms        0ms        0ms        0ms        0ms        0ms        0ms
    policy_group_1           18.03s     18.03s    72.00us        0ms        0ms        0ms        0ms        0ms        0ms        0ms        0ms        0ms
    

  • Event Log messages may be seen like below:

Mon Aug 09 21:18:15 +0800 [node1: kernel: Nblade.cifsLockStateMismatch:debug]: params: {'pMessage': 'Leaked share lock?'}
Mon Aug 09 21:19:38 +0800 [node1: kernel: Nblade_CifsOperationTimedOut_1:error]: params: {'remoteBladeID': 
'12345678-9012-3456-7890-123456789012 (node1)', 'cmdRestartCnt': '0', 'vserverId': '4', 'commandName': 'SMB2_COM_CREATE',
'isQosEnabled': 'QoS_disabled', 'lastSpinNpError': 'SPINNP_NO_FO_ERROR', 'localIpAddress': '10.1.2.3', 'lastCsmError': 
'CSM_TIMEOUT', 'clientIpAddress': '10.2.3.4', 'suspensionCnt': '7', 'dsId': '0'}