nimbus service will not start - no controller process seen - libnim_shared.so file too short

Products

• DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

UIM robot is down and will not start.

We have tried using "nimit" and "service nimbus stop/start" but it won't come up.

By running ps -ef we can see "nimbus /opt/nimsoft" process running but no controller, hub, etc.

If we manually cd to /opt/nimsoft/robot and run ./controller we get an error :

# ./controller

./controller: error while loading shared libraries: /opt/nimsoft/lib/libnim_shared.so; file too short

The file /opt/nimsoft/lib/libnim_shared.so is 0 bytes.

Environment

Release : 20.3

Component : UIM - ROBOT

Cause

bad robot_update

Resolution

The /robot/lib files are replaced on upgrade of the robot or redeployment of robot_update.  Sometimes this can fail e.g. due to a restart at the wrong time, intermittent network failure, etc. and when the robot next restarts, it tries to replace the file and ends up failing.

The solution is to reinstall the robot or copy the /robot/lib/ folder from another robot and this should allow the robot to start.  Then you should deploy the latest version of robot_update to ensure that the libraries and binaries match.