Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-11-03 Thread Pubudu Dissanayake
On Mon, Nov 3, 2014 at 11:05 AM, Kishanthan Thangarajah kishant...@wso2.com
 wrote:

If that's the case, we can go ahead with graceful server shutdown restart
 approach.


​
​+1 for the approach​
 ​



 Also write some integration test cases using HZ Map EntryListener as-well.
 With this, you can test what was the exact values updated for the session
 replication scenario from different nodes.


​Indeed, It will be very useful. BTW I've implemented Hazelcast
EntryListener for hazelcast-map-Persistence feature [1] take a look at it
as well.  ​
​[1] https://github.com/pubudu08/hazelcat-map-persistence​​
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-11-02 Thread Kishanthan Thangarajah
If that's the case, we can go ahead with graceful server shutdown restart
approach.

Also write some integration test cases using HZ Map EntryListener as-well.
With this, you can test what was the exact values updated for the session
replication scenario from different nodes.

On Sat, Nov 1, 2014 at 1:19 AM, Krishantha Samaraweera krishan...@wso2.com
wrote:



 On Fri, Oct 31, 2014 at 6:47 PM, Pubudu Dissanayake pubu...@wso2.com
 wrote:

 ​Thanks for the reply, I'll update our redmine [1] accordingly ​

 On Mon, Oct 27, 2014 at 7:33 AM, Krishantha Samaraweera 
 krishan...@wso2.com wrote:

 Running four nodes in single server is not recommended at integration
 tests phase, so killing a worker node at integration tests run time is not
 possible (Even no admin services exposed in worker nodes to forcefully
 shutdown the instance) I'm not sure whether upcoming operation center will
 facilitate this kind of functionality.

 What is the recommended approach to test my scenario ?


 At the moment this is not possible to automate with current infrastructure
 (and lack of testability of clusters) You can write your own puppet scripts
 etc to do developer testing. But you won't be able to integrate it with our
 build servers.

 You can use the RM item I shared here as a reference for automation test
 availability.

 Thanks,
 Krishantha.




 Thanks,
 Krishantha

 ​[1] https://redmine.wso2.com/issues/2816​ ​

 --
 *Pubudu Dissanayake*
  Software Engineer

 WSO2 Inc.; http://wso2.com
 lean.enterprise.middleware
 Mob: + 94 775 503 304

 Blog: http://geekdetected.wordpress.com/
 Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
 http://lk.linkedin.com/in/pubududissanayake
 Flickr : https://www.flickr.com/photos/pubudufx/
 Twitter: https://twitter.com/GeekInAction




 --
 Krishantha Samaraweera
 Senior Technical Lead - Test Automation
 Mobile: +94 77 7759918
 WSO2, Inc.; http://wso2.com/
 lean . enterprise . middlewear.




-- 
*Kishanthan Thangarajah*
Senior Software Engineer,
Platform Technologies Team,
WSO2, Inc.
lean.enterprise.middleware

Mobile - +94773426635
Blog - *http://kishanthan.wordpress.com http://kishanthan.wordpress.com*
Twitter - *http://twitter.com/kishanthan http://twitter.com/kishanthan*
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-10-31 Thread Pubudu Dissanayake
​Thanks for the reply, I'll update our redmine [1] accordingly ​

On Mon, Oct 27, 2014 at 7:33 AM, Krishantha Samaraweera krishan...@wso2.com
 wrote:

Running four nodes in single server is not recommended at integration tests
 phase, so killing a worker node at integration tests run time is not
 possible (Even no admin services exposed in worker nodes to forcefully
 shutdown the instance) I'm not sure whether upcoming operation center will
 facilitate this kind of functionality.

 What is the recommended approach to test my scenario ?


 Thanks,
 Krishantha

 ​[1] https://redmine.wso2.com/issues/2816​ ​

-- 
*Pubudu Dissanayake*
 Software Engineer

WSO2 Inc.; http://wso2.com
lean.enterprise.middleware
Mob: + 94 775 503 304

Blog: http://geekdetected.wordpress.com/
Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
http://lk.linkedin.com/in/pubududissanayake
Flickr : https://www.flickr.com/photos/pubudufx/
Twitter: https://twitter.com/GeekInAction
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-10-31 Thread Jayanga Dissanayake
Hi All,

In WSO2 Operations Center we are currently supporting only
forceful/graceful restart/shutdown. To do this, we use the ServerAdmin
service. This is similar to restarting/shutting down the server with
management console.
If you want to restart/shutdown a server you can use the ServerAdmin
service.

IMO, If you want to simulate a instant exit of a server (kill -9) in s
remote machine, you have to have some external agent program in that
particular machine or use remote shell.

Thanks,
Jayanga.

*Jayanga Dissanayake*
Senior Software Engineer
WSO2 Inc. - http://wso2.com/
lean . enterprise . middleware
email: jaya...@wso2.com
mobile: +94772207259

On Fri, Oct 31, 2014 at 6:47 PM, Pubudu Dissanayake pubu...@wso2.com
wrote:

 ​Thanks for the reply, I'll update our redmine [1] accordingly ​

 On Mon, Oct 27, 2014 at 7:33 AM, Krishantha Samaraweera 
 krishan...@wso2.com wrote:

 Running four nodes in single server is not recommended at integration
 tests phase, so killing a worker node at integration tests run time is not
 possible (Even no admin services exposed in worker nodes to forcefully
 shutdown the instance) I'm not sure whether upcoming operation center will
 facilitate this kind of functionality.

 What is the recommended approach to test my scenario ?


 Thanks,
 Krishantha

 ​[1] https://redmine.wso2.com/issues/2816​ ​

 --
 *Pubudu Dissanayake*
  Software Engineer

 WSO2 Inc.; http://wso2.com
 lean.enterprise.middleware
 Mob: + 94 775 503 304

 Blog: http://geekdetected.wordpress.com/
 Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
 http://lk.linkedin.com/in/pubududissanayake
 Flickr : https://www.flickr.com/photos/pubudufx/
 Twitter: https://twitter.com/GeekInAction


 ___
 Dev mailing list
 Dev@wso2.org
 http://wso2.org/cgi-bin/mailman/listinfo/dev


___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-10-31 Thread Krishantha Samaraweera
On Fri, Oct 31, 2014 at 6:47 PM, Pubudu Dissanayake pubu...@wso2.com
wrote:

 ​Thanks for the reply, I'll update our redmine [1] accordingly ​

 On Mon, Oct 27, 2014 at 7:33 AM, Krishantha Samaraweera 
 krishan...@wso2.com wrote:

 Running four nodes in single server is not recommended at integration
 tests phase, so killing a worker node at integration tests run time is not
 possible (Even no admin services exposed in worker nodes to forcefully
 shutdown the instance) I'm not sure whether upcoming operation center will
 facilitate this kind of functionality.

 What is the recommended approach to test my scenario ?


At the moment this is not possible to automate with current infrastructure
(and lack of testability of clusters) You can write your own puppet scripts
etc to do developer testing. But you won't be able to integrate it with our
build servers.

You can use the RM item I shared here as a reference for automation test
availability.

Thanks,
Krishantha.




 Thanks,
 Krishantha

 ​[1] https://redmine.wso2.com/issues/2816​ ​

 --
 *Pubudu Dissanayake*
  Software Engineer

 WSO2 Inc.; http://wso2.com
 lean.enterprise.middleware
 Mob: + 94 775 503 304

 Blog: http://geekdetected.wordpress.com/
 Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
 http://lk.linkedin.com/in/pubududissanayake
 Flickr : https://www.flickr.com/photos/pubudufx/
 Twitter: https://twitter.com/GeekInAction




-- 
Krishantha Samaraweera
Senior Technical Lead - Test Automation
Mobile: +94 77 7759918
WSO2, Inc.; http://wso2.com/
lean . enterprise . middlewear.
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


Re: [Dev] Integration tests for Hazelcast Session Replication feature

2014-10-26 Thread Krishantha Samaraweera
Running four nodes in single server is not recommended at integration tests
phase, so killing a worker node at integration tests run time is not
possible (Even no admin services exposed in worker nodes to forcefully
shutdown the instance) I'm not sure whether upcoming operation center will
facilitate this kind of functionality.

Proposed deployment framework can be used to achieve this. However work is
still under progress. You can find the relevant readmine at [1]

Please find relevant mail discussions on worker node management in
automated tests at [2]

[1] https://redmine.wso2.com/issues/2083
[2] http://mail.wso2.org/mailarchive/dev/2013-November/024977.html

Thanks,
Krishantha

On Fri, Oct 24, 2014 at 6:21 PM, Pubudu Dissanayake pubu...@wso2.com
wrote:

 Hi Folks,

 I've implemented Hazelcast based Tomcat Session Replication feature for
 Appserver [1 https://redmine.wso2.com/issues/2816].I have few concerns
 over integration tests for this feature.

 Following is the basic cluster scenario that I have to be written as
 integration tests.
 Environment : *WSO2 ELB, 1-Mgt+2-Worker (WSO2 AS) *

 Basic scenario ( Assume 'x' webapp deployed in Mgt node and enabled
 dep-sync with two worker nodes. )

 Say that we have above mentioned setup. Since our ELB is session aware, it
 will send the requests to the same node based on the seesionID. And when
 that node goes off line, it will route the requests to the other node. But
 if we have enabled replication, this routing should work seamlessly where
 the session, which was replicated and will be used by the second node.

- Start Mgt and WorkerNode-1 -- Change session values at WorkerNode-1
-- Start the WorkerNode-2 -- Perform kill -9 operation on WorkerNode-1
after Workernode-2 complete it's start up operation. Then check whether
those session values changes are reflected at WorkerNode-2.


-  After Change session values at WorkerNode-2 while WorkerNode-1 off
line -- Start the WorkerNode-1 -- Perform kill -9 operation on
WorkerNode-2 after Workernode-1 complete it's start up operation. Then
check whether those session values changes are reflected at WorkerNode-1.

 This routing should work seamlessly where session attributes are
 replicated among all nodes.

 Basically I have to write an integration test to evaluate the feature.
 Please guide me through the process how should I carried out and any
 references to help me further to achieve this integration test? Let me know
 if you need any further suggestions.

 - Pubudu

 [1] https://redmine.wso2.com/issues/2816


 --
 *Pubudu Dissanayake*
  Software Engineer

 WSO2 Inc.; http://wso2.com
 lean.enterprise.middleware
 Mob: + 94 775 503 304

 Blog: http://geekdetected.wordpress.com/
 Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
 http://lk.linkedin.com/in/pubududissanayake
 Flickr : https://www.flickr.com/photos/pubudufx/
 Twitter: https://twitter.com/GeekInAction




-- 
Krishantha Samaraweera
Senior Technical Lead - Test Automation
Mobile: +94 77 7759918
WSO2, Inc.; http://wso2.com/
lean . enterprise . middlewear.
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev


[Dev] Integration tests for Hazelcast Session Replication feature

2014-10-24 Thread Pubudu Dissanayake
Hi Folks,

I've implemented Hazelcast based Tomcat Session Replication feature for
Appserver [1 https://redmine.wso2.com/issues/2816].I have few concerns
over integration tests for this feature.

Following is the basic cluster scenario that I have to be written as
integration tests.
Environment : *WSO2 ELB, 1-Mgt+2-Worker (WSO2 AS) *

Basic scenario ( Assume 'x' webapp deployed in Mgt node and enabled
dep-sync with two worker nodes. )

Say that we have above mentioned setup. Since our ELB is session aware, it
will send the requests to the same node based on the seesionID. And when
that node goes off line, it will route the requests to the other node. But
if we have enabled replication, this routing should work seamlessly where
the session, which was replicated and will be used by the second node.

   - Start Mgt and WorkerNode-1 -- Change session values at WorkerNode-1
   -- Start the WorkerNode-2 -- Perform kill -9 operation on WorkerNode-1
   after Workernode-2 complete it's start up operation. Then check whether
   those session values changes are reflected at WorkerNode-2.


   -  After Change session values at WorkerNode-2 while WorkerNode-1 off
   line -- Start the WorkerNode-1 -- Perform kill -9 operation on
   WorkerNode-2 after Workernode-1 complete it's start up operation. Then
   check whether those session values changes are reflected at WorkerNode-1.

This routing should work seamlessly where session attributes are replicated
among all nodes.

Basically I have to write an integration test to evaluate the feature.
Please guide me through the process how should I carried out and any
references to help me further to achieve this integration test? Let me know
if you need any further suggestions.

- Pubudu

[1] https://redmine.wso2.com/issues/2816


-- 
*Pubudu Dissanayake*
 Software Engineer

WSO2 Inc.; http://wso2.com
lean.enterprise.middleware
Mob: + 94 775 503 304

Blog: http://geekdetected.wordpress.com/
Linkedin: *http://lk.linkedin.com/in/pubududissanayake*
http://lk.linkedin.com/in/pubududissanayake
Flickr : https://www.flickr.com/photos/pubudufx/
Twitter: https://twitter.com/GeekInAction
___
Dev mailing list
Dev@wso2.org
http://wso2.org/cgi-bin/mailman/listinfo/dev