I am experimenting with the cluster sharding activator, but am having lots 
of issues with it. I have tried updating the activator to 2.3.1, but to no 
avail (and other issues show up, such as described here: 
https://www.assembla.com/spaces/akka/simple_planner#/ticket:3967).

Problems noticed so far:

1) 100% of the time, the activator sends a lot of messages to the 
ClusterSystem deadLetters on startup of the seed node. Here is one example:

[INFO] [03/31/2014 09:37:00.654] 
[ClusterSystem-akka.actor.default-dispatcher-2] 
[akka://ClusterSystem/deadLetters] Message 
[akka.cluster.InternalClusterAction$InitJoin$] from 
Actor[akka://ClusterSystem/system/cluster/core/daemon/firstSeedNodeProcess#-438400827]
 
to Actor[akka://ClusterSystem/deadLetters] was not delivered. [1] dead 
letters encountered. This logging can be turned off or adjusted with 
configuration settings 'akka.log-dead-letters' and 
'akka.log-dead-letters-during-shutdown'.
[ ... many more akka.cluster.InternalClusterAction$InitJoin$ messages ... ]
[INFO] [03/31/2014 09:37:05.518] 
[ClusterSystem-akka.actor.default-dispatcher-14] 
[akka://ClusterSystem/system/cluster/core/daemon/firstSeedNodeProcess] 
Message [akka.dispatch.sysmsg.Terminate] from 
Actor[akka://ClusterSystem/system/cluster/core/daemon/firstSeedNodeProcess#-438400827]
 
to 
Actor[akka://ClusterSystem/system/cluster/core/daemon/firstSeedNodeProcess#-438400827]
 
was not delivered. [6] dead letters encountered. This logging can be turned 
off or adjusted with configuration settings 'akka.log-dead-letters' and 
'akka.log-dead-letters-during-shutdown'.
[... JOINING and Up message...]
[INFO] [03/31/2014 09:37:06.516] 
[ClusterSystem-akka.actor.default-dispatcher-16] 
[akka://ClusterSystem/user/sharding/AuthorListingCoordinator/singleton] 
Message [akka.contrib.pattern.ShardCoordinator$Internal$Register] from 
Actor[akka://ClusterSystem/user/sharding/AuthorListing#1471529820] to 
Actor[akka://ClusterSystem/user/sharding/AuthorListingCoordinator/singleton] 
was not delivered. [7] dead letters encountered. This logging can be turned 
off or adjusted with configuration settings 'akka.log-dead-letters' and 
'akka.log-dead-letters-during-shutdown'.
[INFO] [03/31/2014 09:37:06.516] 
[ClusterSystem-akka.actor.default-dispatcher-16] 
[akka://ClusterSystem/user/sharding/PostCoordinator/singleton] Message 
[akka.contrib.pattern.ShardCoordinator$Internal$Register] from 
Actor[akka://ClusterSystem/user/sharding/Post#589187748] to 
Actor[akka://ClusterSystem/user/sharding/PostCoordinator/singleton] was not 
delivered. [8] dead letters encountered. This logging can be turned off or 
adjusted with configuration settings 'akka.log-dead-letters' and 
'akka.log-dead-letters-during-shutdown'.

2) Using the default shared LevelDB journal configuration, sometimes (but 
not always) when the Bot node is started, the seed node goes nuts:

[INFO] [03/31/2014 09:46:00.768] 
[ClusterSystem-akka.actor.default-dispatcher-3] 
[Cluster(akka://ClusterSystem)] Cluster Node 
[akka.tcp://[email protected]:2551] - Leader is moving node 
[akka.tcp://[email protected]:50327] to [Up]
Uncaught error from thread 
[ClusterSystem-akka.remote.default-remote-dispatcher-24] shutting down JVM 
since 'akka.jvm-exit-on-fatal-error' is enabled for 
ActorSystem[ClusterSystem]
Uncaught error from thread [ClusterSystem-akka.actor.default-dispatcher-17] 
shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for 
ActorSystem[ClusterSystem]
Uncaught error from thread [ClusterSystem-akka.actor.default-dispatcher-28] 
shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for 
ActorSystem[ClusterSystem]
Uncaught error from thread [ClusterSystem-akka.actor.default-dispatcher-29] 
shutting down JVM since 'akka.jvm-exit-on-fatal-error' is enabled for 
ActorSystem[ClusterSystem]
[...keeps going forever...]
^CJava HotSpot(TM) 64-Bit Server VM warning: Exception 
java.lang.OutOfMemoryError occurred dispatching signal SIGINT to handler- 
the VM may need to be forcibly terminated

3) When it is working, the shared leveldb journal seems to work reasonably 
well (except for the SPOF on the first node). However, when I change to 
either one of the MongoDB replicated journals in contrib, when testing 
various combinations of node failures, things go nuts with 
duplicatekeyexceptions (looping infinitely), OutOfMemoryError's, and other 
weirdness. I know these are early implementations but the similarity of the 
failures when using the two different journal implementations makes me 
think the problems may not be with the journal implementations, but with 
akka-persistence instead.

4) When restarting the Bot node, there are lots of WARNings about unknown 
UIDs (the following message keeps repeating for Bots that have been shut 
down -- i.e. the node never appears to be actually removed from the 
cluster, even after the entire cluster is restarted):

[WARN] [03/31/2014 10:01:40.280] 
[ClusterSystem-akka.remote.default-remote-dispatcher-5] [Remoting] 
Association to [akka.tcp://[email protected]:50327] with unknown UID 
is reported as quarantined, but address cannot be quarantined without 
knowing the UID, gating instead for 5000 ms.

Has anyone else done any experimentation with akka-cluster-sharding?

Regards,
Raman

-- 
>>>>>>>>>>      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 [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/akka-user.
For more options, visit https://groups.google.com/d/optout.

Reply via email to