Redundancy configuration - A case study

What is needed for “redundancy”?

For a meaningful redundancy circuit, you need 2 CS141s that work together: To make things easier, in this tutorial, we assume a standard case where 2 UPS systems are in use, each connected to different circuits, and then together they ensure power supply through their outputs, for example, with a transfer switch.

The transfer switch will only take notice of a problem if one of the two UPS systems shuts down or is completely shut down for maintenance, and, if necessary, will switch automatically to the other power source. Logically, a transfer switch has no information about the operating state of each UPS or how much remaining runtime is available:

If we assume that Input A is active and Input UPS 1 fails, the transfer switch would switch to Input B as soon as UPS 1 shuts down.

The transfer switch would do the same if UPS Input 1 and 2 fail simultaneously, and, depending on the load, UPS 2 would then take over the load generated by the server until the batteries are exhausted.

According to this logic, the server must be shut down in time before both UPS systems can no longer provide enough emergency power.

Where to configure the redundancy function?

Start with the RCCMD Client – you need to configure several entries:

1. Which device is allowed to send shutdown command?

2. Setting up the redundancy

3. Connection test

4. CS141 configuration

5. How to abort a partial shutdown request?

1. Which device is allowed to send shutdown command?

At the RCCMD client, open options and click on „Connections“:

You need to define a valid sender – in this case, you need to enter the IP addresses of both CS141 WEBMANAGER. Once entered, only these two CS141 are accepted RCCMD command senders.

Note: The DHCP mode may change IP addresses

If running the CS141 WEBMANAGER in DHCP mode, it is possible that the DHCP server will assign new IP addresses. Consequently, the RCCMD client will reject commands because of a changed IP address! To avoid this problem, ensure that the IP addresses of both CS141 WEBMANAGER are fixed. This also applies in reverse – if the DHCP server assigns a new IP address to the RCCMD client, the RCCMD shutdown commands will fail…

2. Setting up the redundancy

After setting up the IP addresses, it is time to configure the redundancy function. To start, click on „Redundancy“:

First, enable the redundancy function, and then select the IP addresses that will form the redundancy group:

How to calculate the redundancy level?

The redundancy level follows a simple assumption:

With two UPS systems, there are always 2 CS141 WEBMANAGERs, one of which will inevitably send a shutdown command first. Which of the two WEBMANAGERs that is depends on the individual state of charge of the batteries, the load, the power failure, etc.

However, the shutdown is suppressed until all other WEBMANAGERs in this list have also sent a shutdown command. This results in the redundancy level being mathematically N + 1, where N is generally the first WEBMANAGER to send a shutdown command, while “+1” is the number of additional WEBMANAGERs that must send a shutdown command in this group.

This results in the following values:

0 → No other CS141 WEBMANAGER is in charge of sending a shutdown signal.

1 → One other WEBMANAGER must send a shutdown command.

Since we only have two UPS systems, you should set the redundancy to 1.

What if the second CS141 is no longer accessible, for example due to a total failure of the switch?

RCCMD accepts the shutdown command with reservation: If one of the two WEBMANAGERs sends a shutdown command, the redundancy function automatically checks whether all other WEBMANAGERs selected for redundancy are available and whether communication with the respective UPS is functioning correctly:

If this condition is not met, the shutdown will be executed because there can obviously be no further shutdown command.

3. Connection test

In the last step, you should check whether RCCMD can reach and query both CS141 WEBMANAGERs. To do this, click “Run alive check now…” under Heartbeats in Test UPS Connections. There are two possibilities:

OK → The CS141 is configured well and reachable.

Not OK → The CS141 or network is not configured well or not reachable.

• Check if Ports 5769 / 6003 TCP are available.
• Check the communication certificate.
• Check if TLS is ON / OFF on both endpoints.
• Check if the CS141 UNMS / RCCMD trap services are active.
• Check if firewalls or virus scan solutions block the communication attempts.

4. CS141 configuration

RCCMD is fully configured and ready to carry out a redundancy shutdown scenario.

Now, both CS141s must be configured identically. Log in to the CS141 WEBMANAGER and carry out the following configuration:

Enable UNMS & RCCMD Trap Services

This step is required to communicate with the CS141 via Port 5769 TCP and query the operation state. As long as this service is OFF, RCCMD will show „NOT OK“ as a UPS alive check result.

Check TLS Settings

This must be identical in the CS141 and RCCMD program (found under Options>Connections), otherwise communication itself will fail.

Define the RCCMD Shutdown Jobs

At the CS141 UPS events „Powerfail“, add the job „RCCMD Shutdown“

Host:

Please enter the target of the job, which is the IP address of the operating system or computer on which the RCCMD client is installed.

Port:

The default port is 6003 TCP – Make sure that RCCMD and CS141 are sending on the same port or listening for an incoming signal.

Timing:

When the CS141 is triggered to send an RCCMD shutdown job. In this example, the job is sent when a power failure is active (Event Powerfail) and the UPS reports that there are 600 seconds of runtime available.

Please note, as both CS141 WEBMANAGERs must send an individually triggered shutdown signal, you need to carry out this configuration on both CS141 WEBMANAGER. RCCMD generally waits for both shutdown commands.

5. How to abort a partial shutdown request?

Let’s recall this drawing. It is realistic that UPS 1 has a power failure that UPS 2 has never been affected by, since they are two separate circuits.

With the current configuration, RCCMD can suppress the shutdown if both UPS systems were otherwise available but cannot independently determine whether the power failure at UPS 1 has been resolved.

However, RCCMD keeps track of which UPS (or which CS141 WEBMANAGER) has already issued a shutdown command, and as soon as all UPS systems have done so, RCCMD will become active and shut down the server.

To meet this, the CS141 offers a special command that revokes its shutdown command:

RCCMD will then delete the shutdown command accordingly and fall back into normal operation. Since the counter event of Power Fail is Power Restored, simply define a corresponding counter job: At Power Restored, select the job “RCCMD Execute”.

Job:

Select „RCCMD EXECUTE“

Host:

Please enter the target for this job. Enter the IP address of the RCCMD client you want to reach.

Port:

The default port is 6003 TCP – Make sure that RCCMD and CS141 are sending on the same port or listening for an incoming signal.

Command:

Use the command „WAKEUP” in the command window – with this command, the CS141 will withdraw the RCCMD shutdown command.

Timing:

When the CS141 is triggered to send an RCCMD shutdown job. In this example, the job is sent when a power failure is active (Event Powerfail) and the UPS reports that there are 600 seconds of runtime available.

Special case:

When the CS141 detects a power failure, sends a shutdown command, and the UPS turns off to save the batteries

For now, the configuration is set up to react to end the power failure before the UPS shuts down. However, if the power failure lasts so long that the UPS decides to shut down completely to protect the batteries from deep discharge, then the next status when the UPS is started up would not be “Power Restored” because this status message is only issued by the UPS when a power failure has ended and the UPS is still running.

When the UPS is restarted, the first event that triggers the CS141 WEBMANAGER is the “UPSMAN started” event, signaling that communication between the CS141 and the UPS has been successfully established and the CS141 is operational.

To ensure that the counter is reset on the respective RCCMD client in both cases, you should also assign the “RCCMD Execute” job with the WAKEUP command to start at UPSMAN started.