PF40 Communication Error

Im after some advice, on a machine we have PF40's and 40P's with 22-COMM-E cards, these plug into a Stratix 5700 switch which is on VLAN 820 on 10.8.20.xx this includes the PLC, drives, and switch etc. from this switch it is connected to a stratix 8000 (distributed switch) which is then connected back to a router on fiber, both the 8000 and 5700 switches are setup with smart ports set as switch for automation, and the router setup to route, the problem im getting is that i had random comms loss (F071) maybe 5 or 6 times a day at random intervals, i found that the OEM had not setup the gateway address on the drives, after entering this it then reduced this to maybe once a day maybe even less, i then removed the uplink cable which went from the 5700 to the 8000 and the faults have gone, has anyone experienced this before? nothing is setup to communicate with these drives across the network, its just a basic backbone network with no PC's servers etc or no comms across different subnets etc.

help!

Rob

Are any errors on the ports being logged on the Device Manager Alert Log? What ports are the devices in question plugged into? What is the firmware of the Stratix 5700 and 8000?

Hi thanks for the reply.

The PLC, VFD's and HMI are in ports 1-9 the uplink cable is in port 19 all on VLAN 83 and the config port for the switch is 19, this is set to VLAN 1

i have looked in the log but nothing appears, i have had some warnings in the switch notification area saying:

portFa1/1: This port is blocked by Spanning Tree Protocol (STP)

this warning appears the same for ports 1 to 9.

the 8000's are the latest firmware revision 15.0(2) the 5700 is 15.0(2)

Do you have any smartports configured? Are these direct connections to the devices? Can you confirm that both switches are set to the same stp?

On the web page:
Vlans --> Advanced

8000 will have a check mark for RSTP. 5700 will have a drop down menu.

on some of the switches smart ports are configured and some they are not but that doesnt appear to make any difference.

yes all switches are set to RSTP

7 connections are direct and 1 is to an unmanaged switch located in the ancillary panel on the machine, i have set this port as multiport automation device.

one other thing, we have four of these machines and on two of them if i unplug the uplink and then plug it back in that port light is amber for a few seconds then flashes green no problems. but on two others if i unplug then plug back in all of the active port lights go amber for 10 seconds or so then change back to green, why does this happen? all switches appear to be setup identical.

That is an issue with spanning-tree. If I had to guess you don't have smartports configured on those ports. RSTP/MST is very dependent on smartport roles because the smartports include a command in the macro that allows the port to be either unaffected when a network event happens or to return comms almost instantly. My suggestion:

Assign all ports smartports
or
Change the switch mode to PVST. This can be done here:
Connect to the webpage
Click the help button at the top left
Click support
Click show information commonly needed by field service
Click Configure
In the command box type spanning-tree mode pvst ; hit enter
In the command box type do wr ; hit enter. It should read Building Configuration...OK
Repeat for both switches.

I think I prefer not using smart ports and just changing the STP to the one you suggest. Should this be applied to every switch on the network? I have a 8300, and two 8000's in fiber ring. With the Gi ports as switch for automation. And from these switches I have copper down links (star) to maybe 12 stratix 5700's and a couple of 8000 located in the machines electrical panels. Located around the plant.

Each of these switches have a unique VLAN ID and IP address...

10.plantnumber.machinenumber.node

What I did have the other day was from what I gathered one of the stratix 8000's rebooted as 6 of our machines (these are standalone bits of kit arranged to form a continuos production line) they each have an uplink to that 8000 switch and each machine stopped, comms errors on all 6 HMI's and all the drives went out on comms error. After maybe 30-60seconds comms ok and drives reset. So what ever happened it made the electrical panel switches stop communicating as all the HMI's, PLC's and drives etc are standalone and do not communicated with other pieces of kit.

These don't have smart ports set so I'm really hoping that if I change the STP this might stop it happening, for now I disconnected the uplinks to all panels till this is sorted.

Our IT guy had a look at the config of the 8300 and two 8000's and he said that there was a lot of config for what we are trying to achieve and he mentioned about why do we need DHCP so I disabled this on all switches not sure if this could be hogging any resources even though all devices have static IP address.

Regards

Rob

The DHCP shouldn't have been enabled by default unless express setup was in use. Either way It is more than likely not resource issue, these switches can handle a lot of traffic.

I would recommended putting PVST on all your swithces. I am 99% certain it will solve your issue.

What about the smart ports for the fiber should I leave them without smart ports enabled and change them to PVST as well?

Our IT guy said we should factory default the switches and start again from scratch using CLI not the GUI I have been using. I think I will try your suggestion first though tomorrow.

Will keep you updated

I would keep the smartports.

The advantage of using the GUI is that it applies the quality of service mappings automatically which is where most of the configuration you see is coming from.

Right you mean go for PVST as well as keep the smart ports on all. If I just change the STP first off and don't configure all the smart ports on the 5700's etc I just the 3 8k's then if this does sort the comms issue then I can take my time and setup the smart ports on the 5700's??

Right, PVST will work with or without the smartports.

I was also getting notifications of ports been blocked due to STP but no loops existed. Hopefully that will disappear too??

You mentioned you had a ring - that would cause a port to be blocked.