There is no default segmentation resolver. If there’s a pause, or network
glitch, you’ll likely end up with two clusters.
> On 28 Sep 2021, at 10:52, 38797715 <38797...@qq.com> wrote:
>
> No, it's completely the default implementation.
>
> I understand that if the cluster has only two nodes, no
No, it's completely the default implementation.
I understand that if the cluster has only two nodes, node1 (Coordinator)
and node2, and if two nodes are disconnected due to long GC, node2
should be shut down due to split brain, but it doesn't seem to be so now.
AFAIK, if the cluster is 3 node
Did you implement a network segmentation plugin?
> On 28 Sep 2021, at 10:38, 38797715 <38797...@qq.com> wrote:
>
> My question is, does the current behavior seem to be inconsistent with the
> default SegmentationPolicy = STOP?
>
My question is, does the current behavior seem to be inconsistent with
the default SegmentationPolicy = STOP?
No, a cluster doesn’t need at least three nodes. According to your logs you
experienced a “long JVM pause.” If a node stops responding, it gets kicked out
of the cluster. It’s likely you need to tune your garbage collection, though
there are other possible causes of long pauses too.
> On 28 Sep
Any feedback?
Hi team,
There is a cluster with 2 server nodes and 1 client node.
It can be seen from the attached log that the two server nodes are
disconnected from about 20:18:15.
However, it is strange that the node with IP 10.97.32.53 did not shut
down due to split brain. Instead, it is split into two
