Cluster state and S3 gateway


(James Cook) #1

Here is a gist of my log files: https://gist.github.com/736518

I am using ES 0.13.0, and the gist includes the elastic search properties.

When I use a gateway.type of none, the server node starts up and the cluster
health is GREEN.

When the server node starts up, its cluster health is RED and remains RED,
even after waiting for the status to change to YELLOW.

Here is the code I use to start the server and wait for YELLOW.

server = nodeBuilder()
.settings(ImmutableSettings.settingsBuilder().put(configuration))
.build();
server.start();
getClient().admin().cluster().prepareHealth()
.setWaitForYellowStatus()
.setTimeout("30s")
.execute().actionGet();
ClusterHealthResponse healthResponse = getClient().admin().cluster()
.prepareHealth()
.execute().actionGet();
LOG.info("ES cluster status is " + healthResponse.getStatus()); // RED?

Any help would be welcome.


(Shay Banon) #2

Based on the gist, there were no indices to be created, right?
On Friday, December 10, 2010 at 8:04 PM, James Cook wrote:

Here is a gist of my log files: https://gist.github.com/736518

I am using ES 0.13.0, and the gist includes the elastic search properties.

When I use a gateway.type of none, the server node starts up and the cluster health is GREEN.

When the server node starts up, its cluster health is RED and remains RED, even after waiting for the status to change to YELLOW.

Here is the code I use to start the server and wait for YELLOW.

server = nodeBuilder()
.settings(ImmutableSettings.settingsBuilder().put(configuration))
.build();

server.start();
getClient().admin().cluster().prepareHealth()
.setWaitForYellowStatus()
.setTimeout("30s")
.execute().actionGet();

ClusterHealthResponse healthResponse = getClient().admin().cluster()
.prepareHealth()
.execute().actionGet();

LOG.info("ES cluster status is " + healthResponse.getStatus()); // RED?

Any help would be welcome.


(James Cook) #3

This is a first time run on EC2, so the gateway held no index data.

-- jim

On Sat, Dec 11, 2010 at 3:49 PM, Shay Banon shay.banon@elasticsearch.comwrote:

Based on the gist, there were no indices to be created, right?

On Friday, December 10, 2010 at 8:04 PM, James Cook wrote:

Here is a gist of my log files: https://gist.github.com/736518

I am using ES 0.13.0, and the gist includes the elastic search properties.

When I use a gateway.type of none, the server node starts up and the
cluster health is GREEN.

When the server node starts up, its cluster health is RED and remains RED,
even after waiting for the status to change to YELLOW.

Here is the code I use to start the server and wait for YELLOW.

server = nodeBuilder()
.settings(ImmutableSettings.settingsBuilder().put(configuration))
.build();
server.start();
getClient().admin().cluster().prepareHealth()
.setWaitForYellowStatus()
.setTimeout("30s")
.execute().actionGet();
ClusterHealthResponse healthResponse = getClient().admin().cluster()
.prepareHealth()
.execute().actionGet();
LOG.info("ES cluster status is " + healthResponse.getStatus()); // RED?

Any help would be welcome.


(system) #4