We met this issue too.
Shall we upgrade to 2.4 to resolve this?
On Sunday, July 20, 2014 9:49:08 PM UTC+8, Patrik Nordwall wrote:
>
> Yes, it can be because of the many join/leave cycles and lack of vector
> clock pruning.
>
> I agree that it is strange that the cluster is working after that, be
> not something that is done by akka itself.
>
> Which Akka version are you using?
>
> /Patrik
>
> 15 jul 2014 kl. 11:00 skrev Xingrun CHEN
> >:
>
>
> We have about 100k actors who sent heartbeat to client side evey several
> seconds. (use system.scheduler.schedule
nd us a stack trace of the scheduler thread?
>
>
> On Tue, Jul 15, 2014 at 7:45 AM, Xingrun CHEN > wrote:
>
>> We found that "ClusterSystem-scheduler-1" thread is very busy and consume
>> 50% cpu (but not more than 50%).
>> And we meet 'no
We found that "ClusterSystem-scheduler-1" thread is very busy and consume
50% cpu (but not more than 50%).
And we meet 'node unreachable' issue every day.
Here's our configs:
akka.cluster {
failure-detector {
acceptable-heartbeat-pause = 6 s # default 3 s
threshold = 12.0# def
Thanks, Patrik. It works well!
On Friday, March 7, 2014 1:03:40 AM UTC+8, Patrik Nordwall wrote:
>
> Hi,
>
>
> On Thu, Mar 6, 2014 at 12:09 PM, Xingrun CHEN
> > wrote:
>
>> To use cluster client, I must register service on server nodes
>> as receptionist
To use cluster client, I must register service on server nodes
as receptionist .
But in the sharding cluster, shard region actor will not register it self
as a receptionist. So in order to access sharding entities through cluster
client, I need to register some 'proxy' actors as receptionist in
