I wondered if there is way to drop a node from cluster if a node is down; I
experieced timeout due to the problem node, which slowed down whole
response time.
I like to know the way to drop the node mannually and still keep the
performance even with the data loss for the node with problem.
If you feel the Elasticsearch instance is slow , just stop the process.
If there is a replica for the data that is there in that node , then that
node would be automatically activated as a shard.
If not , that data wont be visible anymore.
I wondered if there is way to drop a node from cluster if a node is down;
I experieced timeout due to the problem node, which slowed down whole
response time.
I like to know the way to drop the node mannually and still keep the
performance even with the data loss for the node with problem.
Thanks for your response.
With replica configruation set to 0, I like to know if there is a way to
take the problem node, which was already down somehow, out of cluster so
that cluster doesn't bother to send index or search requests to that node
even though I would have some data lost in indexing or unsearchable.
Let me put it in different way; I like to know any way to keep performance
when a node is down with no replica; data loss during indexing or having
some data unsearchable is acceptable.
Hopefully I made my question clearer.
Any response will be appreciated.
Thanks,
JS
On Wednesday, July 4, 2012 10:12:28 PM UTC+9, Vineeth Mohan wrote:
What do you mean by drop a node ?
If you feel the Elasticsearch instance is slow , just stop the process.
If there is a replica for the data that is there in that node , then that
node would be automatically activated as a shard.
If not , that data wont be visible anymore.
Thanks
Vineeth
On Wed, Jul 4, 2012 at 5:22 PM, JS wrote:
Hello,
I wondered if there is way to drop a node from cluster if a node is down;
I experieced timeout due to the problem node, which slowed down whole
response time.
I like to know the way to drop the node mannually and still keep the
performance even with the data loss for the node with problem.
AFAIK - When a node goes down , the master will de register that node and
act as if that node never existed and keep going on with its work with the
data it has.
Can you paste some log where a node waits until timeout for the node that
has disappeared.
Thanks for your response.
With replica configruation set to 0, I like to know if there is a way to
take the problem node, which was already down somehow, out of cluster so
that cluster doesn't bother to send index or search requests to that node
even though I would have some data lost in indexing or unsearchable.
Let me put it in different way; I like to know any way to keep performance
when a node is down with no replica; data loss during indexing or having
some data unsearchable is acceptable.
Hopefully I made my question clearer.
Any response will be appreciated.
Thanks,
JS
On Wednesday, July 4, 2012 10:12:28 PM UTC+9, Vineeth Mohan wrote:
What do you mean by drop a node ?
If you feel the Elasticsearch instance is slow , just stop the process.
If there is a replica for the data that is there in that node , then that
node would be automatically activated as a shard.
If not , that data wont be visible anymore.
Thanks
Vineeth
On Wed, Jul 4, 2012 at 5:22 PM, JS wrote:
Hello,
I wondered if there is way to drop a node from cluster if a node is
down; I experieced timeout due to the problem node, which slowed down whole
response time.
I like to know the way to drop the node mannually and still keep the
performance even with the data loss for the node with problem.
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.