Cluster services with a dependency on a service that was recovered undergo recovery themselves even though depend_mode="soft" in a RHEL 6 High Availability cluster with central_processing mode

Solution Verified - Updated -

Issue

  • services with a dependency on another service that just recovered went in a recoverable state themselves even though depend_mode="soft"
May 29 15:17:53 node1 rgmanager[17958]: Marking service service:nfs01 on down member 3 as stopped
May 29 15:17:53 node1 rgmanager[17958]: Starting service:nfst01 on [ 1 2 5 4 6 ]
May 29 15:18:27 node1 rgmanager[17958]: recover on oracledb "Q11" returned 1 (generic error)
May 29 15:18:27 node1 rgmanager[17958]: Stopping service service:orat01
May 29 15:18:27 node2 rgmanager[9697]: recover on oracledb "SQ1" returned 1 (generic error)
May 29 15:18:27 node2 rgmanager[9697]: Stopping service service:ora03
  • We have services that are dependent on another service to start, but operationally oracle do not need it and should not fail when it fails. Since the dependency is soft, why were these affected when the depend service failed?

Environment

  • Red Hat Enterprise Linux (RHEL) 6 with the High Availability Add On
  • rgmanager

    • <rm central_processing="1"> in /etc/cluster/cluster.conf
    • One or more services with a soft service dependency in `/etc/cluster/cluster.conf:
    <service name="service1" ...>
    [...]
    <service name="service2" `depend="service:service1" depend_mode="soft" ...>
    
    • The dependency service (service1 in the example above) has failed or the node it was running on was failed, leading to recovery

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content