How to do fast fault tolerant fail-over between Active / Passive nodes?

How to do fast fault tolerant fail-over between Active / Passive nodes?

book

Article ID: KB0074727

calendar_today

Updated On:

Products Versions
TIBCO ActiveMatrix BusinessWorks 6.x

Description

If an  active appnode fails, another AppNode that has an application in the passive (stand-by) state takes over and starts processing requests.

If the standby node becomes active, the previous active node is gracefully put into standby state. This would mean to stop all event sources on the appnode and wait for all existing process instances to complete.

There's no fixed time interval for the switch here mainly because it depends on various factors like the hardware of the infrastructure setup, the deployment topology, the network latency between the appnodes.

Issue/Introduction

If the active Appnode is down, how long does it take for the stand-by appnode to detect the active appnode is down and take over as the active appnode?

Resolution

To control the time interval in which a node is brought to standby state  a new property introduced at the appspace level

BW Engine Active to Standby Timeout property is introduced bw.engine.group.ats.timeout=60.

BW Engine Active to Standby Timeout. This is an optional property and it specifies:

1) The time to wait (in seconds) before force stopping an application on an appnode that is transitioning from active to standby state. The default value is "60" secs.
2) 0 secs indicates the appnode waits till the application is gracefully stopped.
 

Additional Information

Please refer below document for more details:
https://docs.tibco.com/pub/activematrix_businessworks/6.5.0/doc/html/GUID-13216E27-4B7A-4669-8B26-87A7B90DE9EC.html