Home› AMX User Forum› AMXForums Archive Threads› AMX Hardware

Network Probs, since change switch

John_GloveJohn_Glove Posts: 95
in AMX Hardware
Hi together

Strange Problem:

I have in my home 3 AMX systems(M2M), working without problem. Last Week i change my gigabit network switch with a new gigabit PoE Switch. After that periodically, the M2M connection fails up/down for 1 or 2 seconds.
After a couple of time the connection is lost. The Port's on the switch are still open. Connection with NetLinxStudio fails! Ping ist possible!
BTW: The PoE function is disabled on this Ports!
On the Switch i see some log entry's like:

Memory Log Error
07:25:48 warning %STP-W-PORTSTATUS: g8: STP status Forwarding
07:25:44 Informational %LINK-I-Up: g8
07:25:42 warning %LINK-W-Down: g8

After i change the switch back to the "old" one and boot the AMX System, all works fine!

Do you have a idea?

thanks
john

Comments

  • HedbergHedberg Posts: 671
    try "zeroconfig disable"
    See here:http://www.amxforums.com/showthread.php?7629-NI-3100-Ports%28IR-RS232%29-sometimes-not-working-and-there-is-no-error-message.&p=52927#post52927

    Also, look at TN#544
  • John_GloveJohn_Glove Posts: 95
    Hedberg wrote: »
    try "zeroconfig disable"
    See here:http://www.amxforums.com/showthread.php?7629-NI-3100-Ports%28IR-RS232%29-sometimes-not-working-and-there-is-no-error-message.&p=52927#post52927

    Also, look at TN#544

    Sorry, i have no access to AMX Website. Do you can send me the TechNote?

    thanks
    john
  • viningvining Posts: 4,368
    I put these warnings at the beginning of the main code for systems as a friendly reminder cuz otherwise I will forget.

    #WARN 'TELNET MASTER & "set udp bc rate 0"'
    #WARN 'TELNET MASTER & set "zeroconf disable"'
    #WARN 'TELNET MASTER & "set duet memory 18m"

    Are most folks basically setting these the same?
  • DHawthorneDHawthorne Posts: 4,584
    vining wrote: »
    I put these warnings at the beginning of the main code for systems as a friendly reminder cuz otherwise I will forget.

    #WARN 'TELNET MASTER & "set udp bc rate 0"'
    #WARN 'TELNET MASTER & set "zeroconf disable"'
    #WARN 'TELNET MASTER & "set duet memory 18m"

    Are most folks basically setting these the same?

    The only one I have ever needed to change from default is the Duet memory setting.
  • the8thstthe8thst Posts: 470
    vining wrote: »
    I put these warnings at the beginning of the main code for systems as a friendly reminder cuz otherwise I will forget.

    #WARN 'TELNET MASTER & "set udp bc rate 0"'
    #WARN 'TELNET MASTER & set "zeroconf disable"'
    #WARN 'TELNET MASTER & "set duet memory 18m"

    Are most folks basically setting these the same?

    Duet memory is the only setting I change on every system.

    I have had to log into network switches and set the ports to a fixed data rate (100 full duplex) as I have had some issues with auto negotiation and AMX gear.
  • viningvining Posts: 4,368
    DHawthorne wrote: »
    The only one I have ever needed to change from default is the Duet memory setting.
    & the8thst,

    I change the upd bc rate to 0 cuz it really serves no purpose if the TPs (maybe other devices) aren't set to auto (if memory serves). Since I don't need it AFIK (I once knew for sure) I don't see the need to broadcast to the entire network every 5 sec (default).

    The zeroconf disable was in light of the TN that Hedberg posted about ealier. Another thing I don't need so I disable it. I don't remember what the TN said but it wasn't good so I figured if I didn't need it, why keep it in the background causing senseless chatter and potentially causing problems.
  • HedbergHedberg Posts: 671
    Actually, the TN refers to the UDP and the link to the message referred to a post by rgelling about the zeroconfig thing.

    For John_Glove:

    Issues like this have been popping up here since before I first logged on and that's been a while. There is something about how Netlinx masters do TCP/IP that sometimes causes problems and, to the best of my knowledge, nobody has actually determined what is the actual cause. Sometimes if seems to be related to certain network equipment (in my case, managed Cisco switches). TN 544 (from 2003, so you see these are not new issues) suggested that some of these problems might be related to the way that AMX masters do UDP conflicting with the way other network gear responds to UDP. For a normal installation, there is no need for UDP so setting the UDP broadcast rate to '0' (i.e. turning udp off) won't hurt and it might make the problem go away.

    Vining covers the three things that are strongly suspected to be linked to these problems and if you are doing a standard installation which does not depend on the zero config or the UDP broadcast, there's no reason to not set them up as he suggests every time.
  • AuserAuser Posts: 506
    John_Glove wrote: »
    Last Week i change my gigabit network switch with a new gigabit PoE Switch. After that periodically, the M2M connection fails up/down for 1 or 2 seconds.

    I can't stress strongly enough how important it is to ensure that the firmware on the controllers is up to date when investigating these sorts of network issues. If you have old masters that haven't been updated for a while, this should be the first step.

    Also, how old are the control systems? Are they NI-x000 series or NI-x100 series? There was a production issue with some NI-x000 series devices affecting the Ethernet port that can cause them to behave strangely like this.
  • John_GloveJohn_Glove Posts: 95
    Auser wrote: »
    I can't stress strongly enough how important it is to ensure that the firmware on the controllers is up to date when investigating these sorts of network issues. If you have old masters that haven't been updated for a while, this should be the first step.

    Also, how old are the control systems? Are they NI-x000 series or NI-x100 series? There was a production issue with some NI-x000 series devices affecting the Ethernet port that can cause them to behave strangely like this.

    I have one NI3000 with Fw (3.00.316/1.00.115) and 3 NI700 (3.50.430/1.20.7)

    Is it possible to check by ser. number which NI's are affected?

    thanks
    john
  • the8thstthe8thst Posts: 470
    vining wrote: »
    & the8thst,

    I change the upd bc rate to 0 cuz it really serves no purpose if the TPs (maybe other devices) aren't set to auto (if memory serves). Since I don't need it AFIK (I once knew for sure) I don't see the need to broadcast to the entire network every 5 sec (default).

    The zeroconf disable was in light of the TN that Hedberg posted about ealier. Another thing I don't need so I disable it. I don't remember what the TN said but it wasn't good so I figured if I didn't need it, why keep it in the background causing senseless chatter and potentially causing problems.

    That all makes perfect sense and it seems obvious, but for some reason I just haven't been doing it.

    I just added the 3 warn messages to my default template. Thanks.
  • AuserAuser Posts: 506
    John_Glove wrote: »
    I have one NI3000 with Fw (3.00.316/1.00.115) and 3 NI700 (3.50.430/1.20.7)

    The 3000 is a big suspect for me, without knowing how the URL list is set up between the three. That firmware is from 2005 and is waaaaaay out of date.

    First step = get it updated.

    There may be issues around updating it as it is so far behind.

    - You may need to bring it up to date by updating to one or more interim versions first.
    - You may need to adjust the Duet memory allocation after the firmware is bought up to date.
    - The updated firmware may (highly unlikely) break existing programming.

    As you don't have access to the AMX website I'm guessing that you may not work on these things on a day to day basis. If so, I definitely wouldn't attempt this yourself as there are lots of gotchas. It is not unlikely that the box will get bricked if you attempt to update it yourself.

    John_Glove wrote: »
    Is it possible to check by ser. number which NI's are affected?

    Don't know about serial numbers, but the issue affected Rev. B and earlier boards. Judging by the age of the firmware on your 3000 it could be in the affected group, but I would pursue updating the firmware first.
  • John_GloveJohn_Glove Posts: 95
    Hi
    I'm testing my Masters Step by Step. The Problem exist's only with the NI3000! all other Masters are working with the new Switch.
    So i try to change the Master with a new NI700.

    Thanks for your hint's!

    john
  • viningvining Posts: 4,368
    John_Glove wrote: »
    Hi
    I'm testing my Masters Step by Step. The Problem exist's only with the NI3000! all other Masters are working with the new Switch.
    So i try to change the Master with a new NI700.

    Thanks for your hint's!

    john
    This may be a dumb question but did you try new cables? If short try longer, if long try a midspan switch. Even it works and the cable passed with a LAN tester (cheap continuity type) try another cable and see if things change.
  • John_GloveJohn_Glove Posts: 95
    vining wrote: »
    This may be a dumb question but did you try new cables? If short try longer, if long try a midspan switch. Even it works and the cable passed with a LAN tester (cheap continuity type) try another cable and see if things change.

    i will try and let you know

    thanks
    john
  • a_riot42a_riot42 Posts: 1,624
    John_Glove wrote: »
    On the Switch i see some log entry's like:

    Memory Log Error
    07:25:48 warning %STP-W-PORTSTATUS: g8: STP status Forwarding
    07:25:44 Informational %LINK-I-Up: g8
    07:25:42 warning %LINK-W-Down: g8

    It might be helpful to mention which switches you are using.
    That's a spanning tree protocol warning. Are you using spanning tree? You can disable it and see it that helps and there are no ethernet loops. Beyond that, check what's connected to port 8 and see if cabling or auto connection speed is an issue.
    Paul
  • John_GloveJohn_Glove Posts: 95
    a_riot42 wrote: »
    It might be helpful to mention which switches you are using.
    That's a spanning tree protocol warning. Are you using spanning tree? You can disable it and see it that helps and there are no ethernet loops. Beyond that, check what's connected to port 8 and see if cabling or auto connection speed is an issue.
    Paul
    Hi

    This is a netgear PoE Gigabit Switch.
    I can disable STP and the problem is still there. on port 8 is the NI3000 connected.
    There is only a problem when the NI is connected.

    john
Sign In or Register to comment.