Isn't it because ActorPath has uid cached for some reason? https://github.com/akka/akka/blob/master/akka-actor/src/main/scala/akka/actor/ActorPath.scala#L137
On Monday, August 11, 2014 10:45:42 AM UTC+4, Patrik Nordwall wrote: > > Hi Johan, > > Thanks for creating a small sample and nice logs. From that I can see and > understand what is going on. > > In the second run the coordinator will use the same state as it had from > the first run, i.e. shard 0 is still allocated to the region actor known > from the first run. That specific actor does not exist any longer, and it > is supposed to discover that by watching the actor. It will discover this, > but before it has received the Terminated message it has already replied to > the GetShardHome request and therefore the initial message is routed to the > old non-existing actor. > > Cluster sharding is not designed to prevent message loss. For re-balancing > and passivation it tries hard to not drop messages, but in case of crash > that is out of scope for cluster sharding. Therefore you need to use some > other mechanism on top to ensure that your messages are delivered, e.g. > AtLeastOnceDelivery in Akka Persistence. > > For this startup scenario we could improve the effort by giving the > Terminated messages from the initial watch a chance to be processed before > replying to GetShardHome requests. I have created a ticket for this: > https://github.com/akka/akka/issues/15646 > > Thanks for reporting! > Regards, > Patrik > > > > On Sat, Aug 9, 2014 at 7:05 AM, Johan Siebens <johan....@gmail.com > <javascript:>> wrote: > >> Hello, >> >> I came across this mail thread because we experience the same issue. >> Currently I'm working an a small project that will consume messages from >> kafka and will be processed be an Akka cluster with Sharding. >> According to the documentation it should be possible to send message to >> the ShardRegion as soon you have an ActorRef, messages will be stashed >> until the required ShardRegion is available. >> During test runs we see this behaviour if we start a node for the first >> time, when we restart the node we're having the 'resolve of path sequence >> failed' debug message and message are send to DeadLetter instead of an >> actor in the ShardRegion. >> If we clean up the underlying persistence storage, the node restart >> correctly >> I was able to reproduce this issue with a small setup (see attachements). >> We're using Akka 2.3.4 and test it with default persistence (LevelDB) and >> with MongoDB, in both cases the node is not able to restart correctly. >> >> Regards, >> Johan >> >> >> >> On Thursday, June 26, 2014 10:16:45 AM UTC+2, Patrik Nordwall wrote: >> >>> Thanks for reporting. Please add your stack trace to ticket >>> https://github.com/akka/akka/issues/15440 >>> If you have any hints of how to reproduce that is very welcome, and >>> please add that to the ticket as well. >>> >>> Cheers, >>> Patrik >>> >>> >>> On Thu, Jun 26, 2014 at 9:58 AM, Carsten Saathoff <car...@kodemaniak.de> >>> wrote: >>> >>>> Hi, >>>> >>>> just for the record, I am also seeing the exact same problem with Akka >>>> 2.3.3. I also use the Cluster Sharding, and the events were stored with >>>> Akka 2.3.3 (it's a fresh setup from yesterday). This only happened after >>>> the system was under constant (but not excessively high) load. In between >>>> one of the nodes died two or three times, and then this strange recovery >>>> behaviour started to occur. Currently the whole cluster is basically >>>> unusable. I can try to create a reproducible test case, but that seems not >>>> so easy. >>>> >>>> best >>>> >>>> Carsten >>>> >>>> Am Dienstag, 13. Mai 2014 16:24:52 UTC+2 schrieb Patrik Nordwall: >>>>> >>>>> Thanks for the update. >>>>> >>>>> As I mentioned previously, the problem can be explained by the >>>>> serialization bug in Akka persistence in 2.3.0 and 2.3.1 that was fixed >>>>> in >>>>> 2.3.2. Are you 100% sure that the data was not stored with 2.3.0 or 2.3.1 >>>>> and then later replayed with 2.3.2? >>>>> >>>>> Please create a ticket if you see the problem again, even though you >>>>> don't know exactly how to reproduce it. >>>>> >>>>> Regards, >>>>> Patrik >>>>> >>>>> >>>>> On Tue, May 13, 2014 at 11:21 AM, benoit heinrich < >>>>> benoit....@gmail.com> wrote: >>>>> >>>>>> Hi Patrik, >>>>>> >>>>>> First of all, I'm very sorry to reply so late but I've tried to >>>>>> create a reproducible test case for you and it seems that I can't easily >>>>>> reproduce the conditions which cause that failure. >>>>>> >>>>>> Also, since I'm trying to recreate a reproducible test case, it seems >>>>>> that even my application works and doesn't throw that error anymore. >>>>>> >>>>>> I'm a bit puzzled, because I'm not sure if I should open a new issue >>>>>> as even myself can't reproduce all the time. >>>>>> >>>>>> I'm going to spend more time this week to try to reproduce and once >>>>>> I've got a way to reproduce all the time then I'll open the issue and >>>>>> attach the test case. >>>>>> >>>>>> >>>>>> Thanks for all, and I'll update that post once I get more news. >>>>>> >>>>>> Cheers, >>>>>> /Benoit >>>>>> >>>>>>> -- >>>>>> >>>>>>>>>> Read the docs: http://akka.io/docs/ >>>>>> >>>>>>>>>> Check the FAQ: http://doc.akka.io/docs/akka/c >>>>>> urrent/additional/faq.html >>>>>> >>>>>>>>>> Search the archives: https://groups.google.com/grou >>>>>> p/akka-user >>>>>> --- >>>>>> You received this message because you are subscribed to the Google >>>>>> Groups "Akka User List" group. >>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>> send an email to akka-user+...@googlegroups.com. >>>>>> To post to this group, send email to akka...@googlegroups.com. >>>>>> >>>>>> Visit this group at http://groups.google.com/group/akka-user. >>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> >>>>> Patrik Nordwall >>>>> Typesafe <http://typesafe.com/> - Reactive apps on the JVM >>>>> Twitter: @patriknw >>>>> JOIN US. REGISTER TODAY! <http://www.scaladays.org/> >>>>> Scala <http://www.scaladays.org/> >>>>> Days <http://www.scaladays.org/> >>>>> June 16th-18th, <http://www.scaladays.org/> >>>>> Berlin <http://www.scaladays.org/> >>>>> >>>>> -- >>>> >>>>>>>>>> Read the docs: http://akka.io/docs/ >>>> >>>>>>>>>> Check the FAQ: http://doc.akka.io/docs/akka/ >>>> current/additional/faq.html >>>> >>>>>>>>>> Search the archives: https://groups.google.com/ >>>> group/akka-user >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "Akka User List" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to akka-user+...@googlegroups.com. >>>> To post to this group, send email to akka...@googlegroups.com. >>>> Visit this group at http://groups.google.com/group/akka-user. >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> >>> >>> >>> -- >>> >>> Patrik Nordwall >>> Typesafe <http://typesafe.com/> - Reactive apps on the JVM >>> Twitter: @patriknw >>> >>> -- >> >>>>>>>>>> Read the docs: http://akka.io/docs/ >> >>>>>>>>>> Check the FAQ: >> http://doc.akka.io/docs/akka/current/additional/faq.html >> >>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user >> --- >> You received this message because you are subscribed to the Google Groups >> "Akka User List" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to akka-user+...@googlegroups.com <javascript:>. >> To post to this group, send email to akka...@googlegroups.com >> <javascript:>. >> Visit this group at http://groups.google.com/group/akka-user. >> For more options, visit https://groups.google.com/d/optout. >> > > > > -- > > Patrik Nordwall > Typesafe <http://typesafe.com/> - Reactive apps on the JVM > Twitter: @patriknw > > -- >>>>>>>>>> Read the docs: http://akka.io/docs/ >>>>>>>>>> Check the FAQ: >>>>>>>>>> http://doc.akka.io/docs/akka/current/additional/faq.html >>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user --- You received this message because you are subscribed to the Google Groups "Akka User List" group. To unsubscribe from this group and stop receiving emails from it, send an email to akka-user+unsubscr...@googlegroups.com. To post to this group, send email to akka-user@googlegroups.com. Visit this group at http://groups.google.com/group/akka-user. For more options, visit https://groups.google.com/d/optout.
