Upgrade In An Akka Cluster

How does one go about it? I'm guessing a rolling restart won't work since different nodes would then have different class versions.

Rolling upgrade by restarting the nodes one by one and allowing the to rejoin is the safest way. If you are not sharing classes through Java serialization but using a generic (and extensible/backward compatible) protocol (like Protobuf) then class versions should not be a problem. It is not trivial though and needs to be designed for, thought through and tested in staging before doing it in a production cluster. 
On Wed, Aug 28, 2013 at 8:55 AM, Jan Vincent Liwanag  wrote:
How does one go about it? I'm guessing a rolling restart won't work since different nodes would then have different class versions.
-- 
>>>>>>>>>> Read the docs: http://akka.io/docs/
>>>>>>>>>> Check the FAQ: http://akka.io/faq/
>>>>>>>>>> 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/groups/opt_out.
-- 
Jonas Bonér
Phone: +46 733 777 123
Home: jonasboner.com
Twitter: @jboner

Hi, Who know about akka cluster ? -- Peter Zhang Sent with Sparrow (http://www.sparrowmailapp.com/?sig)
Hi, Who know about akka cluster ?
Hi, Is there a way to broadcast a message across the nodes in a cluster without using an intermediate actor like a mediator actor in distributed pub-sub or a shard region actor in cluster sharding? BR, Vishnu 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
Hi, I'm just in the early stages of migrating a simulation that uses the old Scala Actors to use Akka instead. The application generates a very large pool of stateful actors that inter-communicate with each other. The functionality in the cluster sharding example (the one the creates blog posts) is quite close to what I need. What I haven't discovered yet is an obvious way to let one actor
Hi, What would be the correct procedure to perform an automated rolling upgrade of an akka cluster (running a dozen of singleton and sharding)? Is it enough to: 1) "leave" one node 2) wait a bit 3) restart the node JVM 4) do the same on another node And should the seed nodes be treated differently? Brice Figureau My Blog: http://www.masterzen.fr/ You received this message because
Hi, I'm currently adding a management system to my Akka Cluster which adds a REST API, allowing to introspect some specific parts of the system (basically some important actor states). This cluster is using several ClusterSharding and ClusterSingletons. One thing I'd like to provide is to be able find on which node a specific shard and shard entry is running through this API, along with
All, I would appreciate some valuable ideas from you all. I have a replication setup with 1 master and 4 slaves, I am short on disk space on all the 5 servers so I will move to another partition (SAN). What is, according to you, the smoothest way to move all 5 servers from For all 5 servers copying data from the local disk to the new partition takes about half an hour. I was thinking
All, I would appreciate some valuable ideas from you all. I have a replication setup with 1 master and 4 slaves, I am short on disk space on all the 5 servers so I will move to another partition (SAN). What is, according to you, the smoothest way to move all 5 servers from For all 5 servers copying data from the local disk to the new partition takes about half an hour. I was thinking
Dear Hakkers. I'm working hard now moving the cluster support in Cloudy Akka into Akka OSS. Will be shipped with Akka 2.0 in the summer. Clustering (or remoting or distribution if you want to call it that) will be completely abstracted away into a configuration task. This means that you can take any Akka application that runs only using local actors and configure it (from the outside) to be clustered
Hi, I was exploring at using play2.0 with akka in a fault tolerant high�availability�cluster. I wanted the community's opinion / impressions on this behalf. How easily doable is it etc ---- Regards, Akhil Kodali
Hello! I've read about the testing with akka-multi-node-testkit, however the project I'm working on is based on Maven3. As far as I know, there's no port of the multi-node testkit plugin for Maven. However I guess it could be possible to mock the cluster behavior somehow with the help of akka-testkit. The sample code to interact with cluster (as from examples) is   val cluster = Cluster(context
We need a HA Scheduler on akka cluster。The scheduler is in local system, If the node crash, the scheduler would lost all timer status. How could impliment a cluster HA Scheduler , no mater which node is crash, the schedule timer would be continue?  You received this message because you are subscribed to the Google Groups "Akka User List" group. To unsubscribe from this group and stop receiving
Hello, everyone! I am relatively new to the Akka world, and I am leading a project based on the Akka Cluster toolbox for an undergrad course, that will be presented to professors a few weeks from now. We are using 64 actors that send String messages every 10 milliseconds to another actor chosen randomly. The receiver actor records the received message in a database. But, for some reason
Hello! I have an FSM actor. which takes a number of references to "worker" actors in it's constructor. That works very well if there's a single actor system. Now I want to create another worker actor *in the cluster* and get a reference to this actor in the FSM actor *outside of the cluster*, so it will be able to send messages to the clustered actor and receive responses. I don't need
Hi I have a 2 node akka cluster, which looses connection once every two days. Often when the server i almost idle. We are still on akka 2.2.4, using oracle java 1.8_25. Its a virtual servers running on vmware esx. The application have been running fine for 3 years, but started with this a few weeks ago. Anyone have any ideas on were to start troubleshooting this? My clusterconfig: remote
Hi All, Can we upgrade multi nodes hadoop cluster from Apache hadoop to CDH through available upgrade steps? or do we need some other approach? Regards Jitendra
Hi all, I want to upgrade a 1TB cluster from hadoop 0.20.3 to hadoop 1.0.3. I am interested to know how long does the hdfs upgrade take and in general how long it takes from deploying new versions until the cluster is back to running heavy MapReduce ? I'd also appreciate it if someone could elaborate about how the hdfs file system is actually executed - I read in the book tat data is not replicated
Hi, our architecture is based on the "share-nothing" approach, so we often have many application nodes that don't know anything about each-other. Now I'm in a situation where I'd like to set up a notification system so that different parts of the system can just send a message to a named actor, which in turn sends that out to all connected (WebSocket) clients. However since there's one such
Hi folks, There are three brokers running 0.8-beta1 in our cluster currently. Assume all the topics have six partitions. I am going to add another three brokers to the cluster and upgrade all of them to 0.8. My question is after the cluster is up, will the partition be evenly distributed to all brokers? Thanks. Regards, Libo
Hi, since we upgraded to Akka 2.3.1 (from version 2.3.0) we are experiencing a strange problem with clustering. More precisely, we are in a very simple situation where we have two seed nodes:     cluster {       seed-nodes = [       "akka.tcp:[email protected]:2551",       "akka.tcp:[email protected]:2552"]       auto-down-unreachable-after = 10s