[Imported] [Solved] Modbus Shared Outstation comms issues

[sbeadle]

>>Message imported from previous forum - Category:ClearSCADA Software<<
User: florian, originally posted: 2018-10-25 19:26:09 Id:279
This is a re-posting from the obsoleted (October 2018) "Schneider Electric Telemetry & SCADA" forum.

-------------------------------

**_hardin4019:_**
**_I am wondering if anyone has any experience with Modbus Shared Outstations?_**

**_I have a situation where Channel A is setup to communicate via Satellite and is the primary, Channel B is setup to communicate via Pepwave BRCellular modem and is the secondary which also acts as the Router with the Satellite going into the WAN port on the Pepwave. Heartbeat is turned on, use preferred Channel (A) is turned on with a delay of 60 seconds, and the Inbound Connections is set to "Accept Inbound Connections on Both Lines". Right now I am using CS 2017 R2, and I am connecting to a Red Lion DSPLE000 that supports multiple connections, and I am polling from the DSP as Modbus TCP. I am using the Heartbeat function to check register 40200, which is a floating point containing a pressure. The Red Lion DSP is polling Modbus TCP from a ROC 800._**

**_For testing purposes, I have 2 cellular modems, a Red Lion DSPLE000, a SCADAPack 330, and a Linksys LRT224 dual WAN router, all setup on my desk and I am seeing the same thing. Both cellular modems are on, connected, and have good communications._**

**_I am seeing lots of communications failures on whichever channel is inactive, even preventing the communications line to switch back to the preffered channel after it has come back up. Looking at the event log, I can see that once both channels are connected, the next Heartbeat is happening roughly 30 seconds later, the inactive channel fails, next Heartbeat it reconnects. Other times, the Heartbeat fails 30 seconds after the channel connects, then within 2 seconds of failing, it reconnects without waiting until the next heartbeat._**

**_Is there a specfic delay I shoud adjust to compensate for a delay in response to the Heartbeat?_**

---------------------

**_hardin4019:_**
**_Something I hadn't tried is using the test setup on my desk, is removing the Red Lion DSP out of the picture and just trying to poll data from the SCADAPack 330. This works perfectly fine and I'm not getting the comms failure on the inactive channel. I will continue to investigate further, but it looks like the Red Lion may not be responding fast enough, be keeping connections open, or allowing enough connections._**

----------------

AFarooq:
You might be confusing the shared with the switched setup. Shared allows for both channels to be used like a load balance concept. Switched uses only one with the choice of a prefered option. Maybe you intended to have a swtiched set up and not a shared setup?

--------------

**_hardin4019:_**
**_I figured it out and have been letting it run for a while. It is "Shared Outstation" and "Shared Outstation Set" for sure. My problem ended up being a connection timeout in the Red Lion DSP. It was set to 10 seconds, an the Heartbeat was only happening every 30 seconds. I increased the timeout to 60 seconds, and all of the issues went away._**

--------------

bevanweiss:
Yeah, I was going to say Wireshark logs or it didn't happen 😛

I think the Wireshark logs would have shown the RedLion issuing a TCP RST (drop connection) which would have caused the ClearSCADA to complain.. it wants to be the one doing the dropping.

You might want to do some testing to see how many concurrent connections the RedLion supports however, we have seen in the past some issues where the slave device only handled two connections. On a redundancy swap there may be up to 4 connections still 'present'.. two from the old server, and two from the new server.

If the new server connections are refused then you will get annoying alarms until the old connections timeout.

----------------

**_hardin4019:_**
**_So far, we have the Red Lion for 2 sessions and it hasn't been an issue. The sites we are polling aren't super critical and a few second or a minute outage isn't going to hurt anything._**

**_Since the original post, I seen an actual outage occur with the test setup on my desk. One AT&T cellular modem, and one Verizon Wireless cellular modem. Something happened in the network and we lost Verizon. The test setup showed the loss of the VZW modem and one or two failed attempts to get it back, but overall, I still had a healthy connection that was updating with "Healthy, B Failed" and when the network issues were resolved, the VZW modem on channel B came back and everything was good and the status changed back to "Healthy, Multidrop"._**

**_The two major issues I was having were having the Red Lion have a single ethernet connection using a single gateway addres, but having two physical gateway modems with two different IP addresses. I tried adding an expansion Ethernet card and IP address to the Red Lion, only one modem would sucessfuly respond to ClearSCADA at a time, and it was only the modem whos IP address was programed as the gateway address of the Red Lion. The Red Lion support was contacted and the Red Lion doesn't support multiple gateway addresses, in fact it stops responding all together if you put in a second gateway address._**

**_Once it was decided I needed a Router with Dual/Multi WAN port capability, the gateway of the Red Lion was switched to the IP of the Router, both modems were responding to ClearSCADA with data from the Red Lion, but the secondary channel was only pulling a heartbeat at the restablis interval and would not stay connected. I now know that the restablish interval was longer than the connection timeout of the Red Lion. The Red Lion was dropping the connection to the second modem because of this, then it would restablish the connection on the next poll._**

**_I guess now I am kind of curious, with all that I have learned, if I can remove the Router and get the second Ethernet port on the Red Lion to work._**