Based on the error i bielive there were some communication issues between the worker node and the master IS process node.
Please make sure to update etc/host file with all node host and IP address
test command ping <hostname> from all nodes in the grid to all other nodes in grid, (node1 > node2, and node2 > node1)
Need additional information to understand the issue:
1. Run strace command on IS pmserver process on all nodes on GRID, once strace running start the workflow til it fails, then end strace and collect output
strace -f -tT -s 256 -o strace_pmserver -p <pmserver PID>
2. Check Core file under INFA_HOME/Server/bin and collect stack trace Refer KB: 13652
3. Integration Service log in debug mode.
4. workflow log.
You should find some additional messages in the Integration service log.
'errno 2' can occur when OS is not able to fork a new process(pmdtm process) for the session.
The failure to fork the child process has to be addressed at operating system level.
Depending on type of OS, please check value of 'ulimit -u' (max processes) for OS user (that ran the session) on UNIX to see if it's set to a low value.
If it's Windows, then work with Windows Admin to enable process monitoring on the server and run the session.
In addition to this, the sessions might fail if there is no port available for communication for DTM which is generally used temporarily from the Maximum-Minimum port range configured at the node.
can you also check on the "Max Process" value set at the node?
You will need to check and set the Maximum and Minimum port range value to a higher one if the Max Process value is increased.
Hoe this helps!