Easily dump/restore complete or partial indices to/from compressed JSON files.
Each index entry will be represented by a line:
{"id":"<id>","type":"<type>","source":{<source>}}
You can use query argument to specify ElasticSearch query in JSON format to filter out records you want to backup.
Example usage arguments:
-hosts=http://host07:9200,http://host06:9200 # ES hosts to connect to
-index=my-index # index to dump
-out=my-index-2017 # file prefix to use
-split=100000 # how many records per file (omit for single file, see note below)
-page=5000 # how many results per scroll to process
-query="{\"range\":{\"dateCrawled\":{\"gte\":\"2017-01-01 0:0:0\"}}}" # ES query to use for partial dumps
Note, that records are kept in memory before they are written to file in compressed form, so it is feasible to specify decent split argument to avoid memory exhaustion for really big indices.
Will produce files like:
my-index-2017.00000000.json.gz
my-index-2017.00000001.json.gz
my-index-2017.00000002.json.gz
...
Example usage arguments:
-hosts=http://host07:9200,http://host06:9200 # ES hosts to connect to
-index=my-index # index to restore
-files=index.*.json.gz # glob file mask