Hi Raj, What you provide in GradientOfHealthRequest.xml execution plan is the time period that CEP waits to execute the Siddhi Window Processor. This is not exactly the time period Stratos will wait to detect faulty instances. It is defined by TIME_OUT variable in FaultHandlingWindowProcessor.java in /extensions/cep/stratos-cep-extension/src/main/java/org/apache/stratos/cep/extension. But it works since termination will only occur after CEP execution plan is executed.
Thanks. On Fri, Oct 10, 2014 at 10:34 PM, Manula Chathurika Thantriwatte <[email protected]> wrote: > Hi Raj, > > I don't think this is configurable. Timeout period is hard coded in the > fault handling window processor at the movement [1]. > > [1] > https://github.com/apache/stratos/blob/master/extensions/cep/stratos-cep-extension/src/main/java/org/apache/stratos/cep/extension/FaultHandlingWindowProcessor.java#L220 > > Thanks ! > > On Fri, Oct 10, 2014 at 8:33 PM, Rajkumar Rajaratnam <[email protected]> > wrote: >> >> Hi Manas, >> >> Seems it is configurable. We do not need to build the stratos again. We >> don't need to even restart the stratos. >> >> You can edit the following file in your running stratos setup. >> >> Please edit the following file; >> >> >> <stratos-setup>/apache-stratos-default/repository/deployment/server/executionplans/GradientOfHealthRequest.xml >> >> There you can find the following code block; >> >> <queryExpressions><![CDATA[ >> from health_grad_processsor >> #window.stratos:faultHandling(1min, member_id) select >> cluster_id, network_partition_id, member_id, partition_id >> group by cluster_id, network_partition_id >> insert into fault_message; >> >> As you can see, we are passing 1min as default. So it will run every one >> min. Change it to 2min. The changes will take effects immediately. You do >> not need to even restart the stratos, as it is supporting hot deployment. >> >> Thanks. >> >> On Fri, Oct 10, 2014 at 8:06 PM, Manula Chathurika Thantriwatte >> <[email protected]> wrote: >>> >>> Hi Manas, >>> >>> Ones CEP received stats from VM, it periodically checks for statics from >>> that VM. Once it didn't received stats from that VM within one minute CEP >>> figure it out as a faulty member and it send member faulty event. >>> >>> Thanks ! >>> >>> On Fri, Oct 10, 2014 at 7:31 PM, Rajkumar Rajaratnam <[email protected]> >>> wrote: >>>> >>>> Hi, >>>> >>>> Forgot to mention that CEP will decide the faulty members based on one >>>> minute window. That is, if an active cartridge is not sending any health >>>> stats events to CEP in any one minute period, it will be identified as a >>>> faulty member. >>>> >>>> I guess, we can't tell the time stratos will take to terminate and >>>> re-spawn it. It depends on the IaaS, network etc. >>>> >>>> Thanks. >>>> >>>> On Fri, Oct 10, 2014 at 7:22 PM, Manas Padegaonkar -T (mpadegao - >>>> ARICENT TECHNOLOGIES HOLDINGS LIMITED at Cisco) <[email protected]> wrote: >>>>> >>>>> Thanks Rajkumar, >>>>> >>>>> >>>>> >>>>> Does it means CEP won’t wait for specific defined time & declare that >>>>> cartridge as faulty-vm?? >>>>> >>>>> >>>>> >>>>> Regards >>>>> >>>>> Manas. >>>>> >>>>> >>>>> >>>>> From: Rajkumar Rajaratnam [mailto:[email protected]] >>>>> Sent: Friday, October 10, 2014 7:14 PM >>>>> To: [email protected] >>>>> >>>>> >>>>> Subject: Re: Which Timer in Stratos Triggers Cartridge Deletion When >>>>> its Unreachable >>>>> >>>>> >>>>> >>>>> Hi Manas, >>>>> >>>>> >>>>> >>>>> If the cartridge-vm becomes unreachable, CEP will not get any events >>>>> from cartridge agent (which is running inside cartridge-vm). Then CEP will >>>>> identify it as a faulty member and publish member fault event. Autoscaler >>>>> will listen to this event and terminate the cartridge-vm as soon as it is >>>>> getting the member fault event. >>>>> >>>>> >>>>> >>>>> Thanks. >>>>> >>>>> >>>>> >>>>> On Fri, Oct 10, 2014 at 7:06 PM, Manas Padegaonkar -T (mpadegao - >>>>> ARICENT TECHNOLOGIES HOLDINGS LIMITED at Cisco) <[email protected]> >>>>> wrote: >>>>> >>>>> Hi Lasindu, >>>>> >>>>> >>>>> >>>>> The timer you mentioned is for the cartridge to come active after it >>>>> got subscribed >>>>> >>>>> >>>>> >>>>> <!-->member expiry timeout (time waited till activation) in ms<--> >>>>> >>>>> <expiryTimeout>900000</expiryTimeout> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> Lets take an example, I subscribed to cartridge & it comes Active >>>>> within the expiry-timeout mentioned above. Now, after sometime the >>>>> cartridge >>>>> vm becomes unreachable or disconnected than based on which timer, stratos >>>>> will decide to delete the cartridge-vm & re-spawn the new instance of it. >>>>> >>>>> >>>>> >>>>> Regards >>>>> >>>>> Manas >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> From: Lasindu Charith [mailto:[email protected]] >>>>> Sent: Friday, October 10, 2014 6:29 PM >>>>> To: dev >>>>> Subject: Re: Which Timer in Stratos Triggers Cartridge Deletion When >>>>> its Unreachable >>>>> >>>>> >>>>> >>>>> Hi Manas, >>>>> >>>>> >>>>> >>>>> You should change the member 'expiryTimeout' property in >>>>> <STRATOS_HOME>/repository/conf/autoscaler.xml in order to adjust the >>>>> timeout >>>>> value. >>>>> >>>>> >>>>> >>>>> Thanks, >>>>> >>>>> >>>>> >>>>> On Fri, Oct 10, 2014 at 6:18 PM, Manas Padegaonkar -T (mpadegao - >>>>> ARICENT TECHNOLOGIES HOLDINGS LIMITED at Cisco) <[email protected]> >>>>> wrote: >>>>> >>>>> Hi, >>>>> >>>>> >>>>> >>>>> Wanted to know which timer in stratos will trigger the deletion of >>>>> existing Active Cartridge when it becomes unreachable. >>>>> >>>>> >>>>> >>>>> Regards >>>>> >>>>> Manas >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> >>>>> Lasindu Charith >>>>> >>>>> Software Engineer, WSO2 Inc. >>>>> >>>>> Mobile: +94714427192 >>>>> >>>>> Web: blog.lasindu.com >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> >>>>> Rajkumar Rajaratnam >>>>> Software Engineer | WSO2, Inc. >>>>> >>>>> Mobile +94777568639 | +94783498120 >>>> >>>> >>>> >>>> >>>> -- >>>> Rajkumar Rajaratnam >>>> Software Engineer | WSO2, Inc. >>>> Mobile +94777568639 | +94783498120 >>> >>> >>> >>> >>> -- >>> Regards, >>> Manula Chathurika Thantriwatte >>> Software Engineer >>> WSO2 Inc. : http://wso2.com >>> lean . enterprise . middleware >>> >>> email : [email protected] / [email protected] >>> phone : +94 772492511 >>> blog : http://manulachathurika.blogspot.com/ >>> >>> >>> >> >> >> >> -- >> Rajkumar Rajaratnam >> Software Engineer | WSO2, Inc. >> Mobile +94777568639 | +94783498120 > > > > > -- > Regards, > Manula Chathurika Thantriwatte > Software Engineer > WSO2 Inc. : http://wso2.com > lean . enterprise . middleware > > email : [email protected] / [email protected] > phone : +94 772492511 > blog : http://manulachathurika.blogspot.com/ > > > -- Akila Ravihansa Perera Software Engineer, WSO2 Blog: http://ravihansa3000.blogspot.com
