I have a use case where I'd like to duplicate an index and then modify the
duplicate in order to preserve the original in case we need to rollback.
The workflow would be:
Index A exists with alias pointing at it
Index B created by duplicating index A
Index B updated with new and changed records
Alias changed to point at Index B
If issue with index B is discovered, repoint alias back to index A
Does anyone have any ideas on a good way to do this?
Sounds good,
you can either use your primary data source to reindex (if you have any
database for instance), or elasticsearch directly. In that case you would
use the SCAN search type
(Elasticsearch Platform — Find real-time answers at scale | Elastic)
to fetch the previously indexed documents in an optimized way. I wouldn't
do the copy then update, but I would just index the updated copy straight
away, as the update would in fact be a reindex and that means that you
would end up reindexing all your documents twice.
Using aliases is definitely the way to go here.
Good luck!
Cheers
Luca
On Tuesday, October 1, 2013 7:23:14 PM UTC+2, mj wrote:
Hi,
I have a use case where I'd like to duplicate an index and then modify the
duplicate in order to preserve the original in case we need to rollback.
The workflow would be:
Index A exists with alias pointing at it
Index B created by duplicating index A
Index B updated with new and changed records
Alias changed to point at Index B
If issue with index B is discovered, repoint alias back to index A
Does anyone have any ideas on a good way to do this?
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.