Hello Legolas,
The auto apply admin feature of LB configuration, would take care of
this. Any admin changes to the cluster setup (ex: addition/removal of
instance..), being load balanced; would be reflected onto the load
balancer configuration maintained at DAS. The auto apply functionality
at DAS (would cause in background, regeneration of loadbalancer.xml)
would take care of propagating these changes over HTTP to the fronting
WebServer. When you create an instance of http Load Balancer at DAS, it
identifies the WebServer on which LB runs via the WebServer instance
host, port details.
HTTP LB upon receiving such update message from DAS, would reconfigure
itself.
regards
Pankaj
legolas wrote:
> Pankaj,
> Thank you for the reply.
> Using GSM APIs look to be a reasonable way to in order to handle this
> problem. But I have a question:
> 1- how will load balancer understand that a new node is added to the
> system?
>
> Thanks.
>
> pankaj jairath wrote:
>
>> Hello Legolas,
>>
>> You can still achieve this concept within GlassFish v2, by using the
>> Self Management feature.
>> -
>> https://glassfish.dev.java.net/javaee5/selfmanagement/selfmanagementhome.html
>>
>> Primarly what you can do is write your custom MBean which listens for
>> GMS events (lets say failure) and causes another instance to added and
>> started within the same cluster. For this you can use the AMX APIs to do
>> this programmatically OR even set up mail alert rule that can send
>> alerts to sys admin when it finds logs of certain type (in this case GMS
>> failure of another peer node of the cluster). This way administrator can
>> rely on the alert notification and take appropriate steps.
>>
>> - https://glassfish.dev.java.net/javaee5/amx/
>> - https://glassfish.dev.java.net/javaee5/docs/DG/beapd.html
>>
>> On a different note, a more intensive solution is Open High Availability
>> Clusters - Open HA Cluster.
>> - http://opensolaris.org/os/community/ha-clusters
>> - http://opensolaris.org/os/community/ha-clusters/ohac/Documentation/
>>
>> This provides HA solution using a combination of hardware and software
>> entities. It provides for the notion of backup nodes as well.
>>
>> regards
>> Pankaj
>>
>> legolas wrote:
>>
>>> Hi
>>> Thank you for you reply.
>>> Is there any way to have such feature?
>>> Any hardware that can do that or any other kind of software which can do
>>> this?
>>>
>>> Thanks
>>>
>>> Kshitiz Saxena wrote:
>>>
>>>
>>>> Hi Legolas,
>>>>
>>>> The load-balancer plugin available with glassfish does not provide any
>>>> such feature of active/passive instances. All instances listed in
>>>> load-balancer.xml are used for load-balancing purposes if they are
>>>> healthy and enabled.
>>>>
>>>> Thanks,
>>>> Kshitiz
>>>>
>>>> legolas wood wrote:
>>>>
>>>>
>>>>> Hi
>>>>> Thank you for reading my post
>>>>> Is there any plan to add active/passive instances support in glassfish
>>>>> load balancer plugin?
>>>>> for example we could have 3 glassfish instance and load balancer
>>>>> plugin utilize two of them at any time, it means that if one of those
>>>>> two active instance goes down, lb plugin send the load to instance
>>>>> number 3, and send some alarm notification about the failed server....
>>>>>
>>>>>
>>>>> thanks
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>>
>>>>>
>>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: users-help_at_glassfish.dev.java.net
>>
>>
>>
>>
>
>