Hi Takahiro,

Please find my comments inline;

Takahiro OGAWA wrote:
> Hi all.
>
> I am testing Load-balance/Failover functions of ESB now,
> and I have found a little question about it.
>
> Both functions allow users to access active servers transparently,
> by setting "disable" flags for downed servers and
> dispatching requests to active servers, that works well.
>   
Cool!
> Then, when the downed server is restored, does ESB become to
> use it automatically?
>   
Yes, but after the suspend duration of the endpoint. Each and every 
endpoint has its own suspend duration and after that duration synapse 
tries to connect to that endpoint and if that also fails, it will be 
again suspended for an another suspended duration and this duration is 
incremental and can be configured.

So the downed server will not appear active soon after restoring but 
after the suspend duration of the endpoint referring to this server, it 
will be activated provided that it is restored and can respond.
> As long as I see, it is not the case...
> If it is not so, is there any settings about polling intervals?
>   
Yeah, I think the above explanation answers this question as well.
> Regards.
>
> P.S.
>
> I found that I can clear "disable" flag from the management
>  console of ESB.
> 1)Open Home > Manage > Mediation > Endopoint
> 2)Edit Failover/ Load-balance Endpoint Group and
>   Save without any change.
>   
This is sort of a hack and you cannot expect the server admin to do this 
always... this happens because when you save the endpoint it creates the 
endpoint definition again which will rest everything related to the 
endpoint configuration.

You don't have to do this because it is properly supported, if you want 
to activate the server ASA it is restored, put a lower value to the 
suspended duration, but be careful because a server which continues to 
be down will affect the performance because synapse will try this 
endpoint first frequently.

Thanks,
Ruwan
> Also ESB can use "disabled" server by listing servers(endpoints)
> repeatedly in a failover endpoint group.(The user have to down
> secondary server after primary server is restored.)
>
> ----
> Takahiro OGAWA
> http://www.brainsellers.com/
>
>
>
> _______________________________________________
> Esb-java-user mailing list
> [email protected]
> https://wso2.org/cgi-bin/mailman/listinfo/esb-java-user
>
>   


_______________________________________________
Esb-java-user mailing list
[email protected]
https://wso2.org/cgi-bin/mailman/listinfo/esb-java-user

Reply via email to