[akka-user] CamelMessage not serializable !?

2014-09-18 Thread Behrad Zari
I was trying to persistent akka-camel CamelMessage objects in my persistent actor, but it complained about it being not serializeable! I'd expect it to be! How can I work around it not to create my own Message class. -- >> Read the docs: http://akka.io/docs/ >> Check

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread √iktor Ҡlang
Yeah, don't access fields before they are initialized, that's recipe for badness. :) On Thu, Sep 18, 2014 at 5:25 PM, Hamilton Turner wrote: > I can confirm my previous message isn't crazy - moving the call to build() > below the line declaring avdName appears to have fixed > the unexpected beha

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread Hamilton Turner
I can confirm my previous message isn't crazy - moving the call to build() below the line declaring avdName appears to have fixed the unexpected behavior, which I take to mean that variables in the constructor are not declared+initialized before other constructor code runs, but are instead only

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread Hamilton Turner
No problem. I've updated the pastie - http://pastie.org/9573588 Only thing I can spot is that build() is being called before the declaration+initialization of name, so perhaps the line `var name:String = ""` is being executed twice. Once before build() to declare the variable, and then once aft

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread √iktor Ҡlang
Could you show me that piece of code? On Thu, Sep 18, 2014 at 4:49 PM, Hamilton Turner wrote: > Ok, heartbeatSchedule makes sense now - there are no threading guarantees > with the future's onSuccess, so I just need to send myself a message > instead. > > I'm still confused about what's happenin

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread Hamilton Turner
Ok, heartbeatSchedule makes sense now - there are no threading guarantees with the future's onSuccess, so I just need to send myself a message instead. I'm still confused about what's happening with variable name. I forgot to include the call to build() (sorry!), it's called in the constructor

Re: [akka-user] Threading issue with Actor and vars

2014-09-18 Thread √iktor Ҡlang
Hi Hamilton, On Thu, Sep 18, 2014 at 1:32 PM, Hamilton Turner wrote: > The below *minimum example* shows some very simple actor code that's not > working as expected. > > I'm setting the value for "name" and for "heartbeatSchedule", but in > postStop they are both set to their initial values > a

[akka-user] Threading issue with Actor and vars

2014-09-18 Thread Hamilton Turner
The below *minimum example* shows some very simple actor code that's not working as expected. I'm setting the value for "name" and for "heartbeatSchedule", but in postStop they are both set to their initial values and not to the values I set. This feels like a threading problem, but it's that

Re: [akka-user] ClusterClient and cluster become unreachable to each other

2014-09-18 Thread Patrik Nordwall
On Thu, Sep 18, 2014 at 2:23 PM, Piyush Mishra wrote: > >Hello Akka hackers > >Is there anybody who is using clusterclient to connect with cluster. I > am using it in my project but akka cluster becomes disjoint to it. What > could be the possible reasons. > >Does cluster client also

Re: [akka-user] Cluster Sharding Supervision Strategy

2014-09-18 Thread Patrik Nordwall
Thanks for reporting! /Patrik On Thu, Sep 18, 2014 at 6:52 PM, Chris Carter wrote: > Thanks Patrik, I've created an issue here: > https://github.com/akka/akka/issues/15943 > > I updated the example to include the RecoveryFailure option, I realized > we're not handling that in our production code

Re: [akka-user] Cluster Sharding Supervision Strategy

2014-09-18 Thread Chris Carter
Thanks Patrik, I've created an issue here: https://github.com/akka/akka/issues/15943 I updated the example to include the RecoveryFailure option, I realized we're not handling that in our production code either. Thanks for the reminder! Chris On Thursday, September 18, 2014 1:11:09 AM UTC-7,

Re: [akka-user] Version of akka-quartz scheduler for akkaVersion = "2.3.6"

2014-09-18 Thread Roland Kuhn
There is nothing magical about the relationship between a scheduler and actor messaging, you can pick any solution that you want and just have the jobs that you schedule send messages as desired. Our reasoning behind not maintaining akka-quartz-scheduler is that other implementors that concentra

Re: [akka-user] Version of akka-quartz scheduler for akkaVersion = "2.3.6"

2014-09-18 Thread Soumya Simanta
Thanks. Is there a replacement for this functionality? Currently I'm using the Akka scheduler. I was planning to move to quartz but looks like the future of quartz is still not clear. > On Sep 18, 2014, at 9:01 AM, Patrik Nordwall > wrote: > > Maintenance of akka-quartz-scheduler has stal

Re: [akka-user] Version of akka-quartz scheduler for akkaVersion = "2.3.6"

2014-09-18 Thread Patrik Nordwall
Maintenance of akka-quartz-scheduler has stalled, and the project is looking for a new maintainer . Regards, Patrik On Wed, Sep 17, 2014 at 8:34 PM, Soumya Simanta wrote: > I'm currently using Akka 2.3.6 and would like to use the Qu

[akka-user] akka-quartz-scheduler is looking for a new maintainer

2014-09-18 Thread Patrik Nordwall
Would you or your organization like to take over the akka-quartz-scheduler project? It will be deleted in a few months unless we find a new owner. -- Patrik Nordwall Typesafe - Reactive apps on the JVM Twitter: @patri

Re: [akka-user] Re: Testing a PersistentActor

2014-09-18 Thread Michael Pisula
Hi Brice, yes, they are pretty similar. What was important to me was the ability to log the messages, as well as cleaning the journal each time the actor system is shut down. If I remember correctly the inmem journal does not do that. I start a new actor system for every test, so that gives me

[akka-user] ClusterClient and cluster become unreachable to each other

2014-09-18 Thread Piyush Mishra
Hello Akka hackers Is there anybody who is using clusterclient to connect with cluster. I am using it in my project but akka cluster becomes disjoint to it. What could be the possible reasons. Does cluster client also need failure detector config? -- >> Read the doc

Re: [akka-user] akka-http protocol considerations

2014-09-18 Thread Patrik Nordwall
Hi Oliver, I don't know what is required by http streaming in node.js. Please let us know if you find that akka-http is missing something for that. Regards, Patrik On Tue, Sep 16, 2014 at 1:22 PM, wrote: > Hello, > > I currently take a look at akka-http. > I'm wondering if there is particular

Re: [akka-user] Akka Cluster 2.3.4 nodes unable to reconnect to cluster

2014-09-18 Thread danny . browning
If you're killing the process, you should send a terminate and use the shutdown hook. sys.addShutdownHook( { val cluster = Cluster(system) cluster.down(cluster.selfAddress) cluster.leave(cluster.selfAddress) system.shutdown() } ) On Thursday, September 11, 2014 11:16:18 AM UTC-6, Joe Won

Re: [akka-user] akka.persistence, snapshots and deleting messages from journal

2014-09-18 Thread danny . browning
Make sure you're receiving the snapshot success message. case SaveSnapshotSuccess(metadata) => { deleteMessages(metadata.sequenceNr, true) } You can also save that metadata in case you want to clean up snapshots. Per Patrik's comment, you may want to verify that some interval has passed befor

Re: [akka-user] akka.persistence, snapshots and deleting messages from journal

2014-09-18 Thread Brian McGee
Hey, I'm implementing a PersistentActor which batches writes for a db. After successfully writing a batch I can't figure out how to determine the latest message number I should use to pass to deleteMessages to prevent the events I just persisted to the db from appearing in the recovery. Any ide

Re: [akka-user] Akka Persistence - issue with Recovery

2014-09-18 Thread Patrik Nordwall
Sounds like a bug. Please create an issue . Thanks for reporting. Regards, Patrik On Wed, Sep 17, 2014 at 7:17 PM, wrote: > I'm seeing a similar issue, where if you handle RecoveryFailure, nothing > happens afterwards. Once the actor enters onRecoveryFailure

Re: [akka-user] Cluster Sharding Supervision Strategy

2014-09-18 Thread Patrik Nordwall
Thank you, Chris! Now I understand the problem. The exception in receiveRecover is caused by event.value.get, and that triggers a restart and doesn't generate the RecoveryFailure message. I agree that we should handle all kinds of exceptions from receiveRecover in the same way. Please create an i