[Hawkular-dev] Is there the feature to provide timegap between ripple restart of WildFly(JBoss EAP) cluster in Hawkular?

[Michael Burman]

Hi,

Silly me, the whole feature exists in RHQ/JON. And as far as I can tell, it's even available in the UI, GroupOperationSchedulesDetailView.

GroupOperationSchedule has:

setHaltOnFailure(boolean) where you can set it to fail.

setExecutionOrder():

     * The order to execute the operations - the first resource in the list is the first one to get its operation
     * invoked. If this is <code>null</code>, all operations can be invoked in any order and may even be done
     * concurrently.

So if you use setExecutionOrder() & setHaltOnFailure() with the GroupOperationSchedule before passing it to the ScheduleManager, you have the feature you requested.

  -  Micke

----- Original Message -----
From: "Michael Burman" <miburman at redhat.com>
To: "Discussions around Hawkular development" <hawkular-dev at lists.jboss.org>
Cc: rhq-users at lists.fedorahosted.org
Sent: Monday, May 2, 2016 3:27:00 PM
Subject: Re: [Hawkular-dev] Is there the feature to provide timegap between ripple restart of WildFly(JBoss EAP) cluster in Hawkular?

Hi,

While you can't find this out of the box in the UI of RHQ/JON, this is easily doable with the RHQ-CLI interface. If you take a look at the samples directory and deploy-to-restart-JBAS.js (or jbossas.js), you can find function function _restartAS7(group) which uses _runOperationSequentially(group, operationName, operationConfig) method. This restarts in sequential order all the AS instances and waits for up to 60 seconds. 

To modify the behavior to do one instance at a time and wait for it to finish and abort if it fails, create a loop around the group object (to find single instance ids) in the _runOperationSequantially, change GroupOperation to single reosurce operation and keep rest of the process intact. That should do all the things you need.

So using either of those scripts as a base provides you the functionality you're looking for. While it's not currently modified to work with Wildfly, that one also is quite simple to add (the resource names / groups are the only things that should need changing).

  -  Micke

----- Original Message -----
From: "Jong Seok Won" <jwon at redhat.com>
To: hawkular-dev at lists.jboss.org
Sent: Friday, April 29, 2016 5:12:16 AM
Subject: [Hawkular-dev] Is there the feature to provide timegap between ripple restart of WildFly(JBoss EAP) cluster in Hawkular?

Hi all~, 

Is there the feature to provide timegap between ripple restart of WildFly(JBoss EAP) cluster in Hawkular? 
If there is no feature, I'm curious there is a roadmap for this feature. 
Actually, this kind of feature is useful in a production environment. 

I've already checked RHQ(JON) however there is no for ripple restart. 
RHQ(JON) provides below in restart operation page of Group menu. 
1. "Execute: in the order specified below 
2. "Halt on Failure?" 


Thanks in advance. 
-- 


Kindly Regards, 

Ted Won 



_______________________________________________
hawkular-dev mailing list
hawkular-dev at lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
_______________________________________________
rhq-users mailing list
rhq-users at lists.fedorahosted.org
https://lists.fedorahosted.org/admin/lists/rhq-users@lists.fedorahosted.org