ElasticSearch and AWS


(Pete Hunt) #1

Hi -

I've set up a 2-node cluster using AutoScaling and Elastic Load
Balancing on Amazon EC2. I have installed the cloud-aws plugin and
configured it as follows:

cloud:
aws:
access_key: [my access key]
secret_key: [my secret key]

discovery:
type: ec2

gateway:
type: s3
s3:
bucket: [my bucket name]

I have created an AMI that starts ElasticSearch with no arguments on
start-up. The problem that I am having is that though searching is
working well, when I hit /_cluster/health, it returns:

{
"status" : "yellow",
"timed_out" : false,
"number_of_nodes" : 1,
"active_primary_shards" : 15,
"active_shards" : 15,
"relocating_shards" : 0
}

Which I believe means that the two nodes are not discovering each
other properly. Am I missing something? The two nodes are identical
AMIs, both running in the same, single security group and the same
availability zone.

Thanks,

Pete


(ppearcy) #2

Just a guess (going to be setting up similar within a few days), but
I'd try explicitly listing the groups and availability_zones listed
here:
http://www.elasticsearch.com/docs/elasticsearch/cloud/aws/ec2_discovery/

Not sure if it is possible to have sensible defaults for these.

Hope this helps,
Paul

On Aug 26, 10:24 am, Pete Hunt floydoph...@gmail.com wrote:

Hi -

I've set up a 2-node cluster using AutoScaling and Elastic Load
Balancing on Amazon EC2. I have installed the cloud-aws plugin and
configured it as follows:

cloud:
aws:
access_key: [my access key]
secret_key: [my secret key]

discovery:
type: ec2

gateway:
type: s3
s3:
bucket: [my bucket name]

I have created an AMI that starts ElasticSearch with no arguments on
start-up. The problem that I am having is that though searching is
working well, when I hit /_cluster/health, it returns:

{
"status" : "yellow",
"timed_out" : false,
"number_of_nodes" : 1,
"active_primary_shards" : 15,
"active_shards" : 15,
"relocating_shards" : 0

}

Which I believe means that the two nodes are not discovering each
other properly. Am I missing something? The two nodes are identical
AMIs, both running in the same, single security group and the same
availability zone.

Thanks,

Pete


(Shay Banon) #3

By default it won't filter by groups. When you execute the cluster state
API, do you see the other nodes?

On Fri, Aug 27, 2010 at 5:23 AM, Paul ppearcy@gmail.com wrote:

Just a guess (going to be setting up similar within a few days), but
I'd try explicitly listing the groups and availability_zones listed
here:
http://www.elasticsearch.com/docs/elasticsearch/cloud/aws/ec2_discovery/

Not sure if it is possible to have sensible defaults for these.

Hope this helps,
Paul

On Aug 26, 10:24 am, Pete Hunt floydoph...@gmail.com wrote:

Hi -

I've set up a 2-node cluster using AutoScaling and Elastic Load
Balancing on Amazon EC2. I have installed the cloud-aws plugin and
configured it as follows:

cloud:
aws:
access_key: [my access key]
secret_key: [my secret key]

discovery:
type: ec2

gateway:
type: s3
s3:
bucket: [my bucket name]

I have created an AMI that starts ElasticSearch with no arguments on
start-up. The problem that I am having is that though searching is
working well, when I hit /_cluster/health, it returns:

{
"status" : "yellow",
"timed_out" : false,
"number_of_nodes" : 1,
"active_primary_shards" : 15,
"active_shards" : 15,
"relocating_shards" : 0

}

Which I believe means that the two nodes are not discovering each
other properly. Am I missing something? The two nodes are identical
AMIs, both running in the same, single security group and the same
availability zone.

Thanks,

Pete


(system) #4