Verifying operation after power line disruption

You can test the MetroCluster configuration's response to the failure of a PDU.

The best practice is for each power supply unit (PSU) in a component to be connected to separate power supplies. If both PSUs are connected to the same power distribution unit (PDU) and an electrical disruption occurs, the site could down or a complete shelf might become unavailable. Failure of one power line is tested to confirm that there is no cabling mismatch that could cause a service disruption.

This test should take about 15 minutes.

This test requires turning off power to all left-hand PDUs and then all right-hand PDUs on all of the racks containing the MetroCluster components.

This procedure has the following expected results:

Errors should be generated as the PDUs are disconnected.
No failover or loss of service should occur.

Turn off the power of the PDUs on the left-hand side of the rack containing the MetroCluster components.

Monitor the result on the console by using the system environment sensors show -state fault and storage shelf show -errors commands.

cluster_A::> system environment sensors show -state fault

Node Sensor State Value/Units Crit-Low Warn-Low Warn-Hi Crit-Hi
---- --------------------- ------ ----------- -------- -------- ------- -------
node_A_1
PSU1 fault
PSU_OFF
PSU1 Pwr In OK fault
FAULT
node_A_2
PSU1 fault
PSU_OFF
PSU1 Pwr In OK fault
FAULT
4 entries were displayed.

cluster_A::> storage shelf show -errors
    Shelf Name: 1.1
     Shelf UID: 50:0a:09:80:03:6c:44:d5
 Serial Number: SHFHU1443000059

Error Type          Description
------------------  ---------------------------
Power               Critical condition is detected in storage shelf power supply unit "1". The unit might fail.Reconnect PSU1

Turn the power back on to the left-hand PDUs.
Make sure that ONTAP clears the error condition.
Repeat the previous steps with the right-hand PDUs.

Give documentation feedback