I'm facing a problem with the ES realtime GET feature. I'm working on a
project where I need to guarantee that once a document is index, all
subsequent reads will find it.
I've created a small test that runs 50 threads and each thread writes a
document and then immediately reads it. The test works ok on a single
node setup but sometimes fails to find the document
on a 2 node setup (approximately 8 times in 100 000 executions). The
failed documents are actually indexed, a manual check finds them
A gist of the main class can be found here - https://gist.github.com/1221648. That class is all you need to run the
test, for convenience a maven project is available on demand.
I'm facing a problem with the ES realtime GET feature. I'm working on a
project where I need to guarantee that once a document is index, all
subsequent reads will find it.
I've created a small test that runs 50 threads and each thread writes a
document and then immediately reads it. The test works ok on a single node
setup but sometimes fails to find the document
on a 2 node setup (approximately 8 times in 100 000 executions). The failed
documents are actually indexed, a manual check finds them
A gist of the main class can be found here - ES problem with realtime get · GitHub. That class is all you need to run the
test, for convenience a maven project is available on demand.
Hi,
I'm facing a problem with the ES realtime GET feature. I'm working
on a project where I need to guarantee that once a document is
index, all subsequent reads will find it.
I've created a small test that runs 50 threads and each thread
writes a document and then immediately reads it. The test works ok
on a single node setup but sometimes fails to find the document
on a 2 node setup (approximately 8 times in 100 000 executions).
The failed documents are actually indexed, a manual check finds them
A gist of the main class can be found here -
https://gist.github.com/1221648. That class is all you need to run
the test, for convenience a maven project is available on demand.
Any ideas, am I missing a setting?
Thanks in advance,
Pavel
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.