search cancel

Job in Event Wait with "Directory do not exist" in Ressource Log

book

Article ID: 241045

calendar_today

Updated On:

Products

CA Automic Dollar Universe

Issue/Introduction

Suddenly, some resources stop being detected by the EEP on a node, so the Uproc with the conditioned on the File Resource remains in Event Wait until it finishes with Time Overrun.

The example below shows the history trace of the Job and a listing of the folder where the file resource is found (done while being logged as root):

In the resource log, we have the following error that explains the root cause:

2022-05-01 23:49:09 | FLUX_G4AC.00000000 | 0009035 | Scan Number XX | No file matching filter. Directory [/fic/g4adp00/reception_sinc/g4ac] do not exist.

We could confirm that the user univa could not access this folder (/fic/g4adp00/reception_sinc/g4ac) with a simple cd command.

Cause

Permissions issue: the user univa (administrator of Dollar Universe / user starting the uxeepsrv process) could not access the folder where the File Resource was pointing to (/fic/g4adp00/reception_sinc/g4ac).

It could be seen while doing a ls -latr in the problematic folder, only the user g4adp00 and the group sinc could access the folder /fic/g4adp00/reception_sinc/g4ac

Additionally, a wrong symbolic link existed in /fic/g4adp00/reception_sinc/g4ac/g4ac pointing to /fic/g4adp00/reception_sinc/g4ac that needed to be deleted.

Environment

Release : 6.x

Component : DOLLAR UNIVERSE

OS: Unix

Resolution

Add execution permissions for the user starting the uxeepsrv process (univa) to the folder where the file is created and delete any wrong symbolic link that could be existing inside that folder.