we run a jdbc river once per night for a complete recreation of our index.
The river - once it has been created - will run again after a restart of
elasticsearch cluster (single node cluster). But what we try to achieve is
a one shot river, which isn't being re-run after restart of cluster.
Naturally we can trigger an event (e.g. as a cronjob) which - after the
estimated end of river run - deletes the river again. But this results in
additional efforts.
The easiest strategy may be do delete the river right away after creating
it. But I suppose in this case the running river would be stopped...?
My question is: what is the best way to create a one shot river which is
not being re-run after restart of cluster?
So I tried the oneshot strategy with no success. The river is being rerun
after restart of cluster.
Am Donnerstag, 12. Februar 2015 15:22:27 UTC+1 schrieb Abid Hussain:
Hi all,
we run a jdbc river once per night for a complete recreation of our index.
The river - once it has been created - will run again after a restart of
elasticsearch cluster (single node cluster). But what we try to achieve is
a one shot river, which isn't being re-run after restart of cluster.
Naturally we can trigger an event (e.g. as a cronjob) which - after the
estimated end of river run - deletes the river again. But this results in
additional efforts.
The easiest strategy may be do delete the river right away after creating
it. But I suppose in this case the running river would be stopped...?
My question is: what is the best way to create a one shot river which is
not being re-run after restart of cluster?
Yes, a river instance is restarted by ES automatically at node restart. You
must delete it after work is done in order to free resources.
Just a note, in future versions, I will drop the ES river API, and the JDBC
plugin will manage an internal state, where the number of execution cycles
can be configured.
So I tried the oneshot strategy with no success. The river is being rerun
after restart of cluster.
Am Donnerstag, 12. Februar 2015 15:22:27 UTC+1 schrieb Abid Hussain:
Hi all,
we run a jdbc river once per night for a complete recreation of our index.
The river - once it has been created - will run again after a restart of
elasticsearch cluster (single node cluster). But what we try to achieve is
a one shot river, which isn't being re-run after restart of cluster.
Naturally we can trigger an event (e.g. as a cronjob) which - after the
estimated end of river run - deletes the river again. But this results in
additional efforts.
The easiest strategy may be do delete the river right away after creating
it. But I suppose in this case the running river would be stopped...?
My question is: what is the best way to create a one shot river which is
not being re-run after restart of cluster?
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.