0

We seem to be running out of shards on some of our larger indices on Elasticsearch 1.7.

In order to utilize larger clusters with more nodes, say 20 instead of 10, I'm hoping that we can create new indices with more shards, and somehow copy the data from the old indices.

I realize that we could re-index from the source application but it is a little impractical in that it would take weeks.

Can we create a new index on a new cluster, with more shards, and simply copy the data with elasticdump?

elasticdump \
  --input=http://localhost:9700/myhugeindex \
  --output=http://newhost:9700/myhugeindex \
  --type=mapping
elasticdump \
  --input=http://localhost:9700/myhugeindex \
  --output=http://newhost:9700/myhugeindex \
  --type=data

Once the transfer is complete, we would invalidate and re-index all changes committed after elasticdump was started.

Might it be done this way, or will it, perhaps, only use the original number of shards on the destination cluster?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.