Viewing the RSTP information of MP-X

[Matt_B_NYNJ]

Is there a proper way to view the Ethernet-port information on the MP-X controllers to see which controller is acting as the "divide" in an R.S.T.P. loop or to see which may have a down port? Also, can someone explain to me again how to view and setup R.S.T.P. on the MP-X devices to make sure it is still happening/happening correctly?

Much appreciated.

(Jeffrey Morton​, the resident MP-X I.T. expert)

[BobClark]

I think you may be interested in the Topology Change Alarm. It has to be created. I would guess it shows the a controller involved.

Alarms for MP Series Controllers​

Configuring an MP Series Controller System Alarm​

Regarding configuring RSTP, there's probably more to be known, but have you seen this?

MP Series Controller Settings Properties – Basic Tab​

[Matt_B_NYNJ]

When I try to open the links, I get this page:

Thoughts?

[BillNeville]

Hi Matt,

You're Schneider SSO is logging you into 1.9 Webhelp. Logoff (upper right corner) and login using UN: EcoStruxureBO@gmail.com and PW: !Preview2018. Select "Show prereleased" and in the Product Version select "Next release".

Then type "Alarms for MP Series Controllers" into the search box.

[Matt_B_NYNJ]

Bob Clark​, I was able to resolve the issue I was seeing with RSTP; it was just to set the priority lower on the switch to gaunertee that it becomes root. The issue was that one of my MP-Xs (I think an MP-V-9A) was becoming the root and could not handle it and brought down the network. I thought I recall Jeffrey Morton​ said that the controllers were capable of being root (just not ideal)... Was I mistaken or is this something I should investigate?

[Jeff]

Bob Clark

I'm used to seeing the Topology Change Notifications without configuring them...from the linked article, it does appear they are built-in System alarms:

System Alarms

System alarms in the MP Series controller include the following:

• Program Cyclic Dependency Alarm: Triggered when a number of programs and other application objects are bound so that the bindings for value transfer form a cyclic dependency (mutually dependent loop). In such cases, the value has no definitive starting point.For more information, see MP Series Controller Program Bindings Cyclic Dependency .
• Program Execution Overrun Alarm: Triggered when a program (or multiple programs) is unable to fulfill all operations within the specified task time. Typically, program execution for a given time slot falls more than 10 samples behind before this alarm displays.For more information, see MP Series Controller Program Execution Overrun .
• Port State Change (Link-up/Link-down) Alarm: Indicates that there is a change in the port state.
• Topology Change Alarm: Displays when there is a fault or a broken topology link. This alarm is informational and requires no user intervention since broken links are repaired automatically in MP Series controllers.

[Jeff]

Matt -

Interesting.  Though I briefly tested MP-X without a managed switch, this is not a configuration that we will want to use.  You will need to have a switch in order to form the ring and allow the Automation Server to communicate with it (i.e. you shouldn't just connect VAV1 to VAV2 to VAV3 to VAV1), and it is unsupported / very risky to attempt using an unmanaged switch in a ring (unmanaged switches don't handle Bridge Protocol Data Units, BPDUs, correctly). 

Since you've already got a managed switch, we will want to configure it (or them if you have multiple) to have a higher Priority / Bridge Priority (to win the Root Bridge elections) to ensure correct network structure when there are no faults.  In short, this will cause the ring to software disable an Ethernet port in the middle of the ring when no faults exist.  In this way, you minimize the number of hops to reach each device. For example, with a ring of 39 MP-X, the ring will function as two daisy chains with 19/20 devices.  I am not aware of any time we would want an MP-X to function as root in an active ring.

The MP-X has a Priority hard coded at 32,768, so you will definitely want to set Priority values lower for your managed switch(es).  It can be set in increments of 4,096, so acceptable values could be: 4096, 8192, 12288, etc.  If you have multiple managed switches on the same network, use different Priority values to govern which should be the root (and which will be the root if connection is lost to the other(s)).

You also need to set a few other parameters to ensure your devices handle 40 device rings correctly: Hello Time, Max Age, and Forward Delay.  My Cisco switch's defaults:

Priority: 32,768

Hello Time: 2

Max Age: 20

Forward Delay: 15

Priority needs to decrease numerically to ensure the network is built around your managed switch, Max Age needs to increase to allow our 40 device per ring limit (i.e. 39 MP-X and one managed switch, or X MP-X controllers and 40-X managed switches), and Forward Delay is related to Max Age.

Adjusted values validated to work with our topology limits:

Priority: 4096, 8192, 12288, etc.  Less than 32,768

Hello Time: 2 (the default for my switch was fine)

Max Age: 40 (to support our 40 device limit...40 is the max supported by RSTP)

Forward Delay: 21

Hope this helps!

--Jeff

[BobClark]

Jeffrey Morton wrote:

Bob,

I'm used to seeing the Topology Change Notifications without configuring them...from the linked article, it does appear they are built-in System alarms:

System Alarms

System alarms in the MP Series controller include the following:

• Program Cyclic Dependency Alarm: Triggered when a number of programs and other application objects are bound so that the bindings for value transfer form a cyclic dependency (mutually dependent loop). In such cases, the value has no definitive starting point.For more information, see MP Series Controller Program Bindings Cyclic Dependency .
• Program Execution Overrun Alarm: Triggered when a program (or multiple programs) is unable to fulfill all operations within the specified task time. Typically, program execution for a given time slot falls more than 10 samples behind before this alarm displays.For more information, see MP Series Controller Program Execution Overrun .
• Port State Change (Link-up/Link-down) Alarm: Indicates that there is a change in the port state.
• Topology Change Alarm: Displays when there is a fault or a broken topology link. This alarm is informational and requires no user intervention since broken links are repaired automatically in MP Series controllers.

Thanks for correcting this Jeffrey Morton. Having not done these myself before, I was following the documentation and made a hasty  conclusion based on this section, but it's clear now that theses System Alarms do not have to be created.

[sesa109157_brid]

Hi Jeff,

Great information! I was made aware of this thread and was about to write a reply of my own to Matt.  Just to reinforce a few of the more important points:

• Managed switches tend to have many ports, typically 8 or more.  MPXs only have 2 ports and are meant to be in a Ring, but not serve as the root.  The diagram below shows an AS-P connected to a Managed 8-port switch to create a private IP network with 3 RSTP rings and an access port to configure the managed switch,  trouble shoot the private IP network, and commission the MPXs and AS-P.  Note that the AS-P does not support RSTP itself and relies on the managed switch to do so.  Since each RSTP ring can have a maximum of 39 MPXs, an 8 port switch with 3 rings can support a maximum of 117 MPXs.  Larger managed switches with more ports can support more rings and therefore more MPXs, up to the limit of 200 MPXs per AS-P.

• When you are configuring RSTP priority keep in mind that the lower the numeric value the higher the priority, which can be confusing at first but that is the way the IEEE RSTP (802.1w / 802.1D-2004) specification is written.
• The MPX’s RSTP priority is not configurable and is hard coded to 32768, which is the lowest possible priority.
• To allow a managed switch to assume the root role just set it's priority higher or to a lower numeric value, I tend to user 4096 which is the highest possible priority.
• Note that RSTP is not limited to just rings, you can use multiple managed switches to effectively create a mesh architecture.  However, this is a more advanced topic and I would encourage everyone to focus on RSTP rings first.

Best regards,

John

[Matt_B_NYNJ]

I am still going to test throughout the day to confirm, but it appears that my issue was that I did not properly restart the MP-Xs after enabling RSTP, so it did not actually get enabled. This is why I was seeing issues. Thank you to John Connor​ for helping point this out.

I am having a bit of trouble re-accessing the MP Series Controller Settings Properties – Basic Tab​ page, but I would like to verify if it mentions to restart the MP-X or not; if it does not, I think it will be an important to add.

[allen_thorpe_co]

Does SE have on switch that would be recommended for this?  One that has been tested already and know for absolute that it works 100% of the time.  I am looking for a 8-port size unit.

Thanks,

Allen

[Jeff]

Allen,

I have been told the Network Topology lab in Boston is working on this.  They have evaluated and blessed some models from Cisco, Moxa, Contemporary Controls, and one or two other brands.  I don't have the exhaustive list, but while we wait for it, I can share a couple things:

1. I've been using the Cisco SG300-10 all year with great success.  It's very cost effective.  I created a short video on how to configure with our required RSTP parameters to support 40 device rings (i.e. 39 MP-X plus the managed switch): Video Link : 3516

2. If you want something that's a little more industrial looking, Cisco has the IE-1000 switch ("Industrial Ethernet").  It costs a few times more (between $400 and $600 depending on the model), but it's DIN rail mountable and powered with 24 VDC.  The Network Topology Lab has also extensively tested the IE-1000.

Hope this helps!