Connection Failure to host:[tcp]servername;port;NODELAY;STREAM

SOLVED

Hello,

We are currently running Sage 100 ERP (as application, not service) and have been having issues.

Recently, the users are intermittently receiving an error stating:

Unable to Connect to Server.

The server "servername"is reachable, but it is not accepting requests on the TCP/IP Socket given. Most likely the Application Server Service is NOT running on the server.

Now when this happens, I look at the server and I have the error:

Connection Failure to host:[tcp]servername;port;NODELAY;STREAM

During this time, I've noticed I can not connect remotely to the MAS server. To resolve this problem, it seems that disabling the workstations NIC card and re-enabling it allows us to connect the workstation back to the server. What could this mean?

  • 0

    Check the spawn tasks setting in user maintenance.

  • 0 in reply to BShockley

    Products

    Sage 100 ERP

    Description

    "Connection failure to host: [TCP] <server>:<port number>: NODELAY: STREAM" trying to access Sage 100 Advanced

    Resolution

    The service may be corrupt or another application is using the same port number as Sage 100. To resolve these issues, perform the following:

    To remove corruption from the Sage 100 service, disable and reinstall the service. Then create new files:

    To uninstall the Sage MAS 200 service:

    From the Windows desktop on the Sage 100 server, click the Start button. On the Programs menu, point to Sage Software, and then click Uninstall Application Server Service.

    In the Service field, select the service, and then select the port of the service.

    Click the Remove button. Click Yes to confirm.

    To reinstall the Application Server Service:

    From the Windows desktop on the Sage 100 server, click the Start button. On the Programs menu, point to Sage Software, and then click Application Server Service Setup.

    Select the path to the Sage 100 folder, and click OK to add the service.

    Note: The service is added based on the settings on the Server tab of Application Server Configuration.

    To refresh the Application Server Configuration files and make the connection more stable:

    Stop the Application Server and close the Application Server Configuration application.

    On the Sage 100 server, rename:

    ..\MAS_System\SY_Console.m4t

    ..\MAS_System\SY_Workstation.m4t

    ..\home\lib\_appserv\LOCATE.PVK

    ..\home\lib\_appserv\SESSIONS.PVK

    Restart the Application Server.

    From the Windows desktop on the Sage 100 server, click the Start button. On the Programs menu, point to Sage Software, and then click Application Server Configuration. Note that new LOCATE.PVK and SESSION.PVK files are created.

    Click the Server Tab. Clear the KeepAlives check box.

    Click the Clients Tab. On the Re-Connect list, select None.

    If problem continues, change the port number:

    From the server, set the service to manually start.

    From the Windows desktop on the Sage 100 server, click the Start button. On the Programs menu, point to Sage Software, and then click Application Server Configuration.

    Click the Server tab. In the Sockets field, change the socket number to any unused number between 9,000 and 10,000.

    Click the Start button.

    If the service starts and runs as an application on the new port number, remove the previous service, and create a new one with the new port number.

    Remove the original service using Uninstall Application Server Service.

    Create a new service using Application Server Service Setup

    If the message is occuring for Specific users:

    Have all users exit Sage 100.

    Log on to Sage 100 as Administrator.

    From Administrative Tools, select User Maintenance.

    Select the first user, and click the Preferences tab.

    Select the Spawn Tasks from Application Server check box, and click Accept.

    Proceed for remaining users.

    Click Done.

    Run Sage 100 as an Application rather than as a Service:

    On the Sage 100 server from the Windows Taskbar, open Start > Program Files > Sage Software > Application Server Configuration.

    Click Start

  • 0 in reply to TomTarget
    verified answer

    "NODELAY; STREAM" Long story short, there is an interuption in communcation between the server and the workstation that Sage 100 Advaced is not able to recover from.  It needs to be troubleshot like a network issue.  NIC's, cables, switches, etc..

    BShockley is correct, the reason you see those messages on the server is because you have 'Spawn Tasks From Application server" checked the the User, Profiles tab.  For now uncheck it for everyone and the messages will appear on the workstation that generated the message.  That could help narrow down where the interuption is starting.

  • 0 in reply to TheOtherGuy

    I have seen that sort of thing happen because of "green" NIC cards to go to sleep. It's been a couple of years but I think the fix was to download and install firmware updates to the cards or change some configuration setting.

  • 0 in reply to jimatqsi

    In my experience, this issue is not related to the network cards.   It is an issue of settings on the file server when Sage 100 is run as a service.   There are some other articles that discuss tweaking registry entries on the server, but usually require some experimentation.

    As recently as last week,  I solved it by using turning ON the spawn task from server setting.  

    That being said, sometimes you end up trying a number of different things to resolve it.  You just have to experiment sometimes.

  • 0 in reply to TomTarget

    TomTarget you are correct but  "NODELAY; STREAM" is a generic message that any interuption can cuase. (I've seen it here in a Sage Irvine training room caused by a bad cable)  Since exceeding the heap memory allocation will cause a disconnection it can cause this message.  However, you will almost always see a "connection timed out" message assosiated with the heap memory issue but he didn't mention seeing that error.  

    Saeed S. said restarting the nic on a workstaion would clear the issue up; he didn't mention getting users out of the system. The heap memory issue has a very predictable threshhold.  Once a certain number of taks is opened in a system, no more tasks will open after that.  That threshhold number of task is set in stone, it will be the same until you expand the non-interactive heap in the system registry  You can test what your threshhold is by opening a specific task (S/O order entry) 30 times each  on several workstations until you start getting locked out.  

    • It seems unlikely (Unlikely but not impossible) other users would exit the system every time he restarted the nic. My bet would be on the NIC settings or the NIC itself.  
    • However, running the applicaiton server as an application (not a service) for a test period is the best way to test whether you are exceeding the service heap memory allocation.  If the issue completely goes away while running as an application, you have your answer.
  • 0 in reply to TheOtherGuy

    Interesting insight.   Thanks for the info.

  • 0 in reply to TomTarget

    There was an issue with one of the cables. Once we swapped it out, the issue was resolved. Looks like it was indeed hardware related. Thank yo all for the help.