Conference azur::mcc

6202.0. "FAIL-OVER DETECTION" by UTROP1::CAMPENHOUT_R (Rob (J.R.) van Campenhout) Tue Dec 27 1994 11:35

    Hi,
    
    A BRS Cluster has been implemented at the ABN-AMRO bank here in
    Holland.
    
    One of the requirements for managing this environment is that CRITICAL 
    alerts which are generated by the individual management products on the OMS
    station (PCW, DECMCC, VCS) are passed on to the central NETVIEW
    management console. 
    
    We have written some software which filters and collects the alerts and
    passes them on to the NETVIEW console on the IBM mainframe. The bank 
    however also requires that the comms link between the OMS station and 
    NETVIEW be made redundent just like the BRS configuration.
    
    This means that when the BRS cluster fails-over the alert forwarding
    must also fail-over. In order to do this our software on the secondary
    OMS station needs to be able to detect that a failover has occurred so
    that we can send alerts over the back-up comms link to NETVIEW.
    
    We know that the VCS process is started but not active on the secondary
    OMS station but that does not help in determining whether fail-over has
    occurred.
    
    Our local BRS engineer has been consulted but he could not provide us
    with a conclusive method for detecting fail-over.
    
    Can anyone indicate whether there is a flag or logical set on the OMS
    station which indicates that a fail-over has occurred and how to test
    for it?
    
    We need to have this implemented by feb 1 1995 so any help will be
    appreciated.
    
    Regards,
    
    Rob