Stale Models and alerts after Spectrum fault tolerant fail-over
Article ID: 198302
Updated On:
Products
Issue/Introduction
Our distributed environment has a fault tolerant pair for each landscape. After failover of these landscapes, we sometimes observe a lot of noise with alerts in the queue re-firing, or device models that were deleted reappearing.
Is there something we are not doing, or perhaps a sync we could perform before a controlled failover to reduce this noise?
Environment
Component : Spectrum Core / SpectroSERVER
Cause
The primary SSDB should be synched to the secondary before failover. This happens automatically after scheduled online backups but will have the SSDB from the last OLB time and so will not be completely up to date.
The type of fault tolerance of the secondary will play a part. i.e. a hot secondary will poll constantly for events while inactive so will be ready to take over more quickly and automatically with up to date events.
A cold or warm secondary will need to be synched first.
Resolution
Please choose the right form of secondary for your environment.
Use the online backup subview from the VNMN model in the OneClick console, to synch the primary to the secondary, before a manual failover.
Additional Information
Feedback
