On a Cluster Installation, splSock Socket Files Are Overwritten once a Managed Server Starts (Doc ID 1934083.1)

Last updated on SEPTEMBER 22, 2022

Applies to:

Goal

splSock socket-files are overwritten once a managed server is starting.

Example Replication Steps:

Weblogic Cluster consists of two nodes: node1 for managed server etmSrv01 and node2 for managed server etmSrv02. The two nodes share a common OCFS file system: <DIR>. ETPM is installed under the common OCFS file system, SPLEBASE = <DIR>/etm/<SERVER>, as a result <DIR>/etm/SERVER>/runtime is common for both nodes.  Both SPLService and SPLWeb are deploying and starting properly. 

Assuming AdminServer and NodeManagers are up and running:
1. Startup Manager Server: etmSrv02 on node2 (time: 23:29)
etmSrv02 starts up normally

2. Notice the sockets created under /<DIR>/etm/SERVER>/runtime:
[cissys@node2 bin]$ ls -ltrh /d<DIR>/etm/SERVER>/runtime/ | grep splSock
srwxr----- 1 cissys cisusr    0 Sep 22 23:29 splSock6504
srwxr----- 1 cissys cisusr    0 Sep 22 23:29 splSock6506
srwxr----- 1 cissys cisusr    0 Sep 22 23:29 splSock6507
srwxr----- 1 cissys cisusr    0 Sep 22 23:29 splSock6508
srwxr----- 1 cissys cisusr    0 Sep 22 23:29 splSock6509

3. Startup Manager Server: etmSrv01 on node1 (time: 23:35)
etmSrv01 starts up normally

4. Now, notice that under <DIR>/etm/SERVER>/runtime new sockets have overwritten the previous ones:

[cissys@node1 bin]$ ls -ltrh /data01/etm/ETMTAS/runtime/ | grep splSock
srwxr----- 1 cissys cisusr    0 Sep 22 23:35 splSock6504
srwxr----- 1 cissys cisusr    0 Sep 22 23:35 splSock6506
srwxr----- 1 cissys cisusr    0 Sep 22 23:35 splSock6507
srwxr----- 1 cissys cisusr    0 Sep 22 23:35 splSock6508
srwxr----- 1 cissys cisusr    0 Sep 22 23:35 splSock6509

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.