| Summary | Application taking a long time to start after Orbix services failover |
|---|---|
| Article Number | 18788 |
| Environment | Orbix 6.x Windows |
| Question/Problem Description | Application taking a long time to start after Orbix services failover Slave services take a long time to register after the locator and other services have been promoted to master |
| Clarifying Information | |
| Error Message | |
| Defect/Enhancement Number | |
| Cause | Orbix will try to contact the primary service 5 times before trying the alternate address specified in the IOR. This leads to a delay when the master is unavailable and the server must contact the slave CFR, locator and Name Service. |
| Resolution | The configuration variables that affect this and there defaults are: policies:binding_establishment:backoff_ratio="2"; policies:binding_establishment:initial_iteration_delay = "100"; policies:binding_establishment:max_binding_iterations = "5"; In the case of the connection to the locator the timeline would be something like: ~0ms - first connection attempt 100ms - second attempt 100 200ms - third retry 300 400ms - fourth retry 700 800ms - fifth retry This should only take approximately 1.5 seconds for the connection to be established. However Orbix relies on the OS to report back that the connection cannot be established and if the OS takes a period of time to find this then this time would have to be added to each of the five attempts. |
| Workaround | |
| Notes | |
| Attachment |
| Created date: | 06 September 2011 |
|---|---|
| Last Modified: | 13 February 2013 |
| Last Published: | 26 June 2012 |
| First Published date: | 09 September 2011 |
#Orbix
#KnowledgeDocs
