Stopping the FailSafe service on a node  Inadvertently shutting down the secondary companion

Chapter 14: Configuring Adaptive Server for Failover on SGI IRIX

Recovering from a failed manual prepare_failback

If you inadvertently take the failed resource group offline before issuing prepare_failback, the primary companion’s devices will be removed by FailSafe. This causes an Adaptive Server error 840 on all the primary databases that have been moved to the secondary companion. The databases are marked as not recovered, offline.

This error can be corrected, but requires advanced System Administration.

To recover from a failed manual prepare_failback:

  1. Bring the primary resource group online. See “Bringing the primary resource group online” for more information.

  2. Take the secondary resource group offline, then online. See “Taking the primary companion’s resource group offline” and “Bringing the primary resource group online” but use sec_rg instead of pri_rg for the resource group name.

  3. Allow updates to the system catalogs. Log in to the secondary companion using isql and enter:

    sp_configure “allow updates”, 1
    
  4. Unset the not recovered flag. Log in to the secondary companion using isql and enter:

    update sysdatabases set status = status & ~256 
    where status & 256 = 256
    
  5. Disallow updates to the system catalogs. Log in to the secondary companion using isql and enter:

    sp_configure “allow updates”, 0
    
  6. (This is a repeat of step 2.) Take the secondary resource group offline, then online. See “Taking the primary companion’s resource group offline” and “Bringing the primary resource group online”, but use sec_rg instead of pri_rg for the resource group name.

  7. Confirm that databases are recovered. Log in to the secondary companion using isql, and enter:

    sp_helpdb
    




Copyright © 2005. Sybase Inc. All rights reserved. Inadvertently shutting down the secondary companion

View this book as PDF