camel-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Claus Ibsen <claus.ib...@gmail.com>
Subject Re: Clustered routes in master/slave mode
Date Tue, 07 Jun 2016 14:12:41 GMT
Logged a ticket about the health check API
https://issues.apache.org/jira/browse/CAMEL-10026

On Mon, Jun 6, 2016 at 12:35 PM, Claus Ibsen <claus.ibsen@gmail.com> wrote:
> On Wed, Jun 1, 2016 at 7:34 PM, Matt Pavlovich <mattrpav@gmail.com> wrote:
>> +1
>>
>> Claus-
>>
>> I could see a lot of value in having this feature for file://, ftp:// and
>> several other components as well.
>>
>> Are you thinking Context or Route level or both?
>>
>
> At first it should be route level as that is the most common use case.
> Also its easier to implement as there is one route in the question.
>
> If its context level, then you need to keep track of more, such as all
> routes and figure out when to fail if any of those routes starts to
> have problems. But down the road we should allow both levels.
>
> To aid both we should also look at introducing some health API in
> camel-core, so that makes this easier as well. So you can check for
> health status on context / routes level. And custom components can
> implement custom logic to check their health, such as connecting to a
> FTP server and do a FTP list, or do a HTTP remote call etc.
>
>
>
>> -Matt
>>
>>
>>
>> On 5/28/16 2:52 AM, Claus Ibsen wrote:
>>>
>>> Hi
>>>
>>> We have a few components that offer capability to run a Camel routes
>>> in master/slave mode, where one node is elected as master and only
>>> runs the route at that node, and the other nodes are in slave mode,
>>> and then will failover if needed.
>>>
>>> For example camel-zookeeper, and camel-quartz can do that. And now
>>> also camel-consul. And possible others I have forgot.
>>>
>>> I wonder if we should introduce some API those components
>>> implementations can implement so we would better be able to manage and
>>> know about this. Then we can have JMX information about routes being
>>> in master/slave mode, and you can see for example in a 4 node cluster
>>> that its node-b that is the master.
>>>
>>> node-a route-foo: slave
>>> node-b route-foo: master
>>> node-c route-foo: slave
>>> node-d route-foo: slave
>>>
>>> We have a bit old JIRA ticket about this:
>>> https://issues.apache.org/jira/browse/CAMEL-4454
>>>
>>>
>>>
>>>
>>
>
>
>
> --
> Claus Ibsen
> -----------------
> http://davsclaus.com @davsclaus
> Camel in Action 2: https://www.manning.com/ibsen2



-- 
Claus Ibsen
-----------------
http://davsclaus.com @davsclaus
Camel in Action 2: https://www.manning.com/ibsen2

Mime
View raw message