search cancel

tiebreaker started as second primary scheduler

book

Article ID: 228898

calendar_today

Updated On:

Products

CA Workload Automation AE - Scheduler (AutoSys)

Issue/Introduction

Tie breaker's instance of Autosys (primary and shadow were already active) had a misconfigured config file,  RoleDesignator=1  was the option set on it. 

When the tie breaker was restarted,  this instance somehow did not identify that the Primary scheduler was fully active and instead started up as another primary.  

Cause

Potential racing condition with the HAPollInterval logic. 

Example, say HAPollInterval=60  on primary and shadow, but HAPollInterval=29  on TieBreaker.

Because the tiebreaker has <1/2 of Primary+shadow  AND   its RoleDesignator=1,  during the check it performs for Primary's presence, a racing condition could potentially develop and cause the tie breaker not to recognize the primary properly.

 

Environment

Release : 11.3.6

Component :

Resolution

1) ensure HAPollInterval is same on all the nodes

2) Ensure RoleDesignator value is set properly on all nodes and not have two servers with primary option pointing to the same autosys database