cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Coli <rc...@eventbrite.com>
Subject Re: too many full gc in one node of the cluster
Date Fri, 13 Nov 2015 21:50:44 GMT
On Thu, Nov 12, 2015 at 10:35 PM, Shuo Chen <chenatu2006@gmail.com> wrote:

> We have a small cassandra cluster with 4 nodes for production. All the
> nodes have similar hardware configuration and similar data load. The C*
> version is 1.0.7 (prretty old)
>
> One of the node has much higher cpu usage than others and high full gc
> frequency, but the io of this node is not high and data load of this node
> is even lower. So I have several questions:
>

The simplest explanation is that your 4th node is the only node your
clients ever talk to.

Other simple explanations include "the 4th node owns more of the ring than
other nodes do."

=Rob

Mime
View raw message