Hi,
My client seems not able to connect to elasticsearch cluster (0.19.3).
Here's client log:
DEBUG;[zen.ping.multicast:94][main] [Jester] using group [224.2.2.4], with port [54328], ttl [3], and address [null]
DEBUG;[zen.ping.unicast:94][main] [Jester] using initial hosts [], with concurrent_connects [10]
DEBUG;[elasticsearch.discovery.zen:94][main] [Jester] using ping.timeout [2m]
DEBUG;[discovery.zen.elect:94][main] [Jester] using minimum_master_nodes [-1]
DEBUG;[discovery.zen.fd:94][main] [Jester] [master] uses ping_interval [1s], ping_timeout [30s], ping_retries [3]
DEBUG;[discovery.zen.fd:94][main] [Jester] [node ] uses ping_interval [1s], ping_timeout [30s], ping_retries [3]
WARN;[org.elasticsearch.discovery:114][main] [Jester] waited for 30s and no initial state was set by the discovery
INFO;[search.analyzer.ElasticSearchIndexer:180][main] Search cluster qa-search01
ERROR;[search.analyzer.ElasticSearchIndexer:205][main] org.elasticsearch.discovery.MasterNotDiscoveredException: waited for [30s]
Here's server log:
[TRACE][discovery.zen.ping.multicast] [Proteus] [1] received ping_request from [[Jester][mVQxLCYzSQugPNA4P3T4Yg][inet[/10.3.5.21:9300]]{client=true, data=false}], sending ping_response{target [[Proteus][_G945VyASLyJ2FrxRs_x7g][inet[/10.3.5.109:9300]]], master [[Proteus][_G945VyASLyJ2FrxRs_x7g][inet[/10.3.5.109:9300]]], cluster_name[qa-search01]}
It seems like the server "Proteus" received the ping request from the client "Jester" and responded,
but the client timed out (30 sec).
Is there anywhere else I should look at to investigate what is happening?
It is solved by adding '-Djava.net.preferIPv4Stack=true' to java command that starts the client.
Before adding this, the multicast ping sent by the client had ttl=1, and after adding this, the ttl=3.
I don't know why this was needed, but it works now.
Further question:
how can I change client setting, such as multicast ttl?
I think ttl=3 is the default. How can I change?
Can that be fed into NodeBuilder.settings(Settings)?
If so, will that be ImmutableSettings.settingsBuilder().put("discovery.zen.ping.multicast.ttl", 4)?
How about "discoverty.zen.ping_timeout"? Can it be set by the same way?
On Friday, July 20, 2012 4:09:22 PM UTC-4, arta wrote:
It is solved by adding '-Djava.net.preferIPv4Stack=true' to java command
that
starts the client.
Before adding this, the multicast ping sent by the client had ttl=1, and
after adding this, the ttl=3.
I don't know why this was needed, but it works now.
Further question:
how can I change client setting, such as multicast ttl?
I think ttl=3 is the default. How can I change?
Can that be fed into NodeBuilder.settings(Settings)?
If so, will that be
ImmutableSettings.settingsBuilder().put("discovery.zen.ping.multicast.ttl",
4)?
How about "discoverty.zen.ping_timeout"? Can it be set by the same way?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.