Skip to Content
0
Jul 30, 2020 at 09:11 AM

Error 13 SAP Mobile Platform Version 3.0 SP17 and Memory Allocation Errors

89 Views

Hello Community,

We are looking for advice on how to successfully setup SAP Mobile Platform to connect to a Mobile Cluster.

Based upon guidance from SAP it isn't clear on how the JavaBe should be used to communicate with a mobile JVM cluster. We have 6 nodes in our PPT environment setup to handle SMP Application traffic.

The entries currently in the JavaBe.ini which relate to the configuration are as follows:

naming.provider.url=corbaloc:iiop:qdc2eam103.onelondon.tfl.local:9814
#naming.provider.url=corbaloc:iiop:amis-ppt1.onelondon.tfl.local:9814
maximo.jndi.rootContext=cell/nodes/Node01/servers/MOB1
naming.factory.initial=com.ibm.websphere.naming.WsnInitialContextFactory
maximo.jndi.ejb.accesstoken=ejb/maximo/remote/accesstokenprovider

At the moment using this configuration we infrequently get an error 13 followed by a memory leak. This is exacerbated by multiple connections and ultimately the window server runs out of committed memory and the only solution is to reboot.

We are running V3.0 SP17 without any additional patches.

Examples of errors which occur are:

System.out###Agentry Runtime Worker 10304 Thread#na#######User::initMxSession::system#notboundexception

System.out###Agentry Runtime Worker 10304 Thread#na#######Server::login:: for IMTEST297 caught AgentryException logging in user IMTEST297 - system#notboundexception

.\agentry\netmsg.cpp#1383:ReceivedNetMessage::disconnected(RCP<Grapevine>)

4,3221225477, Thr       7472, EXCEPTION_ACCESS_VIOLATION, 00007FF85B58CAA7, 0, R, 0000000005E00000, ..\agentry\netmsg.cpp#1371:ReceivedNetMessage::disconnected(RCP<Grapevine>)

07/28/2020 08:39:35, 1,         4,2147483649, Thr       7456, Unknown Windows Structured Exception, 00007FF85B58CAA7, 0, ..\agentry\netmsg.cpp#1371:ReceivedNetMessage::disconnected(RCP<Grapevine>)

Our main questions are:

  1. Are the errors related to misconfiguration of the JavaBe.ini
  2. And if so is using a WebSphere cluster supported.
  3. If using a cluster is supported should we be using corba-loc in the naming provider URL with the bootstrap port for each member of the cluster (we have tried this and still got the error)
  4. Or using the deprecated IIOP config:
naming.provider.url=iiop://tdc2eam122.onelondon.tfl.local:9810maximo.jndi.rootContext=cell/nodes/Node01/servers/UI1