Redundant servers with redundant connections on same LAN segment

[Ladrone]

Hello to Everyone,

I'm developing a redundant project with this system architecture:
-WinccOA 3.15 patch 15
-2 Siemens Simatic Server Based on Windows Server 2012 R2 with 2 different ethernet cards
-Redundant Network Connections

At the first attempt i configured the redundant architecture in this way:
-I started from a non-redundant project
-I created Host file with this configuration

/****************************************************************************/
192.168.133.10 TRS-SRV-0001 TRS-SRV-0001-1 #phisical ethernet card 1 of server 1
192.168.133.20 TRS-SRV-0001-2 #phisical ethernet card 2 of server 1

192.168.133.11 TRS-SRV-0002 TRS-SRV-0002-1 #phisical ethernet card 1 of server 2
192.168.133.21 TRS-SRV-0002-2 #phisical ethernet card 2 of server 2
/****************************************************************************/

-I modified project config file with this configuration

/****************************************************************************/
[general]
...
...
...
...
data = "TRS-SRV-0001,TRS-SRV-0001-2$TRS-SRV-0002,TRS-SRV-0002-2"
event = "TRS-SRV-0001,TRS-SRV-0001-2$TRS-SRV-0002,TRS-SRV-0002-2"
[ui]
connectToRedundantHosts = 1
/****************************************************************************/

-I Added Redundancy Manager
-I Added Split mode Manager
-I Added Control Manager : fileSync.ctl -connectToRedundantHosts -num99
-I shared the project folder
-Copy of redundant project in the other server

After this configuration, when i opened the system status panel and started to test the system the redundancy had a strange behavior.
Just one of two redundant connections beetween EV managers lines became green and if I opened the panel of UI gui, I saw that only link 0 of two server turns green.

At the second attempt i changed the LAN segment of the second ethernet card LAN interfaces:

/****************************************************************************/
192.168.133.10 TRS-SRV-0001 TRS-SRV-0001-1 #phisical ethernet card 1 of server 1
192.168.134.20 TRS-SRV-0001-2 #phisical ethernet card 2 of server 1

192.168.133.11 TRS-SRV-0002 TRS-SRV-0002-1 #phisical ethernet card 1 of server 2
192.168.134.21 TRS-SRV-0002-2 #phisical ethernet card 2 of server 2
/****************************************************************************/

And the system worked properly.

My question is, why the redundancy didn't work when the LAN segment of the ethernet cards (1 and 2) of those 2 server are in the same network segment?

I hope i was clear.

Thanks in advance.

[leoknipp]

I have never tested redundant network connections for network interfaces in the same segment and therefore I cannot tell you if it works or not.
To get more detailed information debug flags can be activated.

If it is working with two different segments you can use this configuration.

The config entry connectToRedundantHosts = 1 must not be defined for the UI. A UI automatically connects to both servers.

Best Regards
Leopold Knipp
Senior Support Specialist

[Gertjan van Schijndel]

Leonardo Mattioli wrote:

At the second attempt i changed the LAN segment of the second ethernet card LAN interfaces:
...
And the system worked properly.

If there are multiple NICs in the same LAN segment, then the OS decides over which card the data will be sent. In case of network interruption of the first network(card) the OS might take more time than WinCC OA to switch to the other network. In the worst case the secondary does also not work (because all data was sent before over the primary).