Connectrix B-Series Fabric OS 8.X: Continuous MAPS alerts IT_FLOW

Flow monitors are configured to count the Initiator or Target (IT) flows to a device and as threshold of 16 as per the default moderate policy on the switch is defined. But because of zoning and device activity, the number of flows has surpassed the threshold for the PIDs for which the alerts were triggered.

Below MAPS alerts are seen in errdump:

2018/05/09-00:00:01, [MAPS-1003], 181579, SLOT 1 FID 101, WARNING, MA_Dataplex2, Pid 0x7af700, Condition=ALL_LOCAL_PIDS(IT_FLOW>16), Current Value:[ IT_FLOW,65 IT-Flow(s)], RuleName=defALL_LOCAL_PIDSIT_FLOW_16, Dashboard Category=Fabric Performance Impact.

2018/05/09-00:00:01, [MAPS-1003], 181580, SLOT 1 FID 101, WARNING, MA_Dataplex2, Pid 0x7af780, Condition=ALL_LOCAL_PIDS(IT_FLOW>16), Current Value:[ IT_FLOW,65 IT-Flow(s)], RuleName=defALL_LOCAL_PIDSIT_FLOW_16, Dashboard Category=Fabric Performance Impact.

Use the below command to display the information of zoned devices with the PID which is mentioned in the alerts that are encountered in errdump output:

nszonemember -n    < Displays each local device's online zoned device data including the device PID.

Port: 5/1              Pid: 0x7af780
               Zoned Members: 64 devices
              
Port: 5/3              Pid: 0x7af700
               Zoned Members: 64 devices

Flow monitors have been configured to count the Initiator or Target (IT) flows to a device and as threshold of 16 as per the default moderate policy on the switch is defined. But because of zoning and device activity, the number of flows has surpassed the threshold for the PIDs for which the alerts were triggered.

-IT_FLOW    < The number of zoned online devices that have exceeded the zoned device ratio threshold.
Rules containing the monitoring system (IT_FLOW) are not monitored in versions before Fabric OS 8.0.1.

-MAPS    < Allows monitoring of the zoned device ratio per port.
In Fabric OS 8.0.1 and later, devices can be zoned to other devices to allow communications. When MAPS finds that a port has more than the expected number of devices zoned-in, then it alerts the administrator. 

-Zone    < Configuration can cause back pressure in the following situations:

• If a device is zoned with a disproportionate number of devices
• If a port is allowed to communicate with a disproportionate number of ports

IT_FLOW is not a feature, it monitors traffic, hence it cannot be disabled.  

The workaround would be to tune the policy as suggested below. In this example, we tuned the value from 16 (current value) to 80 (80 is an example). This means that the policy should not trigger an alert until the threshold reaches 80.

Below are the steps that are provided to tune the threshold value from 16 (current value) to 80. (80 is an example, you can set the threshold to a value you select.):

1. #> mapspolicy --clone dflt_moderate_policy -name new_moderate_policy
2. #> mapsrule --clone defALL_LOCAL_PIDSIT_FLOW_16 -rulename newdefALL_LOCAL_PIDSIT_FLOW_80
3. #> mapspolicy --delrule new_moderate_policy -rulename defALL_LOCAL_PIDSIT_FLOW_16
4. #> mapsrule --config newdefALL_LOCAL_PIDSIT_FLOW_80 -op ge -value 80 -policy new_moderate_policy
5. #> mapspolicy --enable new_moderate_policy