We have a 3-node cluster (157,158 and 159), we send 4000messages/s to 158
in 157, sometimes 157 can not receive the heartbeat from 158, but 158 is
already sent the heartbeat. The heartbeat between 159 and 158 is fine, and
157-159 also ok. Our version is akka2.3.10, and hope to be able to solve
We have a 3 nodes cluster (157,158 and 159), we send 4000 messages per
seconds to 158 in 157, sometimes 157 can not receive the heartbeat from
158, but 158 is already sended. The heartbeat between 159 and 158 is fine,
and 157-159 also ok. Our version is akka2.3.10, and we want to solve this
constrained to this role.
>
> --
> Konrad `ktoso` Malawski
> Akka <http://akka.io> @ Lightbend <http://lightbend.com>
>
> On 15 August 2016 at 04:23:26, Yutao Shuai (szq...@gmail.com )
> wrote:
>
>
> <https://lh3.googleusercontent.com/-tZtLqoK9VUY/V7EnCA2ZVKI
Is there a way to reload configuration at run time?
For example:
I defined failure detect parameter
"monitored-by-nr-of-members = 1"
while the cluster consists of 3 nodes.
I want to adjustment
"monitored-by-nr-of-members = 5"
after some new nodes have been joined to the cluster.
Thanks.
--
Does the leader will down himself or another node will become leader? Where
is the code about this in the source files. Thanks much for your help!
--
>> Read the docs: http://akka.io/docs/
>> Check the FAQ:
>>
In a 3-node cluster, all of nodes are seed nodes, if first node start
failed, the whole cluster can't start successful.
--
>> Read the docs: http://akka.io/docs/
>> Check the FAQ:
>> http://doc.akka.io/docs/akka/current/additional/faq.html
>>
Indirect probe described as below. After every T time units, a process
Miselects
a random process from its membership list - say Mj - and sends a *ping* to
it. It then waits from an *ack* fromMj.If it does not receive the *ack* within
the pre-specified timeout, Miindirectly probes Mj by
Can I use the SBR for free?
在 2016年7月22日星期五 UTC+8下午10:02:13,Patrik Nordwall写道:
>
> Yes, it exists for 2.3.x
> fre 22 juli 2016 kl. 13:01 skrev Yutao Shuai <szq...@gmail.com
> >:
>
>> Can I use split brain resolver in version 2.3.10 ?
>>
>> 在 2016年7月22日星期五
ightbend customers and provides ready-made solutions (the linked pages),
> but if you want to go with the open source version, you can implement
> similar strategies yourself.
>
> -Endre
>
> On Fri, Jul 22, 2016 at 11:18 AM, Yutao Shuai <szq...@gmail.com
> > wrote:
>
>
s/akka/rp-16s01p05/scala/split-brain-resolver.html
> or similar.
>
> /Patrik
> fre 22 juli 2016 kl. 09:41 skrev Viktor Klang <viktor...@gmail.com
> >:
>
>> What behavior are you looking to achieve?
>>
>> --
>> Cheers,
>> √
>>
>> O
The result we expected is divide into two cluster, one of the cluster has
three or four nodes and work normally.
在 2016年7月22日星期五 UTC+8下午3:41:32,√写道:
>
> What behavior are you looking to achieve?
>
> --
> Cheers,
> √
>
> On Jul 22, 2016 5:11 AM, "Yutao Shuai&
In a 5-node cluster, hypothesis the nodes is A,B,C,D,E. All nodes monitor
each other, enable auto-down, and we set the node-A random package loss
rate 50%, after a period of time, the cluster
will divide into two or more clusters. How can I fix this problem?
--
>> Read the
In a 5-node cluster, hypothesis the nodes is A,B,C,D,E. All nodes monitor
each other, enable auto-down, and we set the node-A random package loss
rate 50%, after a period of time, all nodes in the cluster will restart.
The result we expect is one or two nodes restart, how can I fix this
Description of the problem:
In a 3-node cluster, all nodes monitor each other, when a connection
between two nodes fails, the cluster divide into 3 cluster, every cluster
have one node.
Example:
For example, In 3-node cluster, the default value of
monitored-by-nr-of-members is 5, and
14 matches
Mail list logo