uxioserv crashes with coredump while saving node settings changes

Products

• CA Automic Dollar Universe

Issue/Introduction

The Dollar Universe services for BVS,SAP, CDJ, DQM, EEP and GSI go down when making a change in Node Setting (values.xml).

The process generates a core file.

Example of errors that can be seen in universe.log:

|WARN |X|BVS|pid=p.t| UXOS_HdlTermProcess       | execution handler: SIGNAL = (6) PID = (XXX) PPID = (1) GPID = (XXX)
|WARN |X|CDJ|pid=p.t| UXOS_HdlTermProcess       | execution handler: SIGNAL = (6) PID = (YYY) PPID = (1) GPID = (YYY)
|ERROR|X|IO |pid=p.t| read_new_network_message  | Network error -2: Connection closed
|ERROR|X|IO |pid=p.t| o_refresh_remote_node_set | Error -1 sending node setting update to engine CDJX¿
|ERROR|X|IO |pid=p.t| u_call_srvio_ver          | Error connecting to CDJX¿X: Connection closed
|WARN |X|DQM|pid=p.t| UXOS_HdlTermProcess       | execution handler: SIGNAL = (6) PID = (ZZZ) PPID = (1) GPID = (ZZZ)
|ERROR|X|IO |pid=p.t| read_new_network_message  | Network error -2: Connection closed
|ERROR|X|IO |pid=p.t| o_refresh_remote_node_set | Error -1 sending node setting update to engine DQMXÄ
|ERROR|X|IO |pid=p.t| u_call_srvio_ver          | Error connecting to DQMXÄ X: Connection closed
|ERROR|X|IO |pid=p.t| read_new_network_message  | Network error -2: Connection closed
|ERROR|X|IO |pid=p.t| o_refresh_remote_node_set | Error -1 sending node setting update to engine EEPX.
|ERROR|X|IO |pid=p.t| u_call_srvio_ver          | Error connecting to EEPX.X: Connection closed

Environment

Release : 6.x

Component : DOLLAR UNIVERSE

Cause

A problem has been fixed where having a large node settings file, for example having a huge amount of Tasks in O_PURGE_SPECIF_JOB_PUR node setting, caused uxioserv process to crash.

For example, with a high number of entries in Maintenance > Specific purge Job list (task name) (O_PURGE_SPECIF_JOB_PUR).

The exact number of Tasks in the list that makes the node crash is not known exactly but seems a combination of (i) the number of Tasks and (ii) the Task name length.

E.g. the node crashes with 200+ Tasks with an average Task name length of 31 characters or 395+ Tasks with a length of 16 characters.

Resolution

Workaround:

Reduce the number of Task in Specific purge and restart the crashed processes when issue occurs

Solution:

Update to a fix version listed below or a newer version if available.

Fix version(s): 
Component: Dollar Universe
Dollar Universe 6.10.101 - Available