Version: 1.6.0
- JDK: 6.0 (or above) but JDK 8 (at least u20 or higher) or JDK 7 (u55) is best recommended
- Hardware: at least 8Gb memory, cpu does not matter much
- Set number of open file to 64000
- Heap size to be less than half of memory but not more than 64 Gb
- Disable memory swap (can use es settings)
- Download & unpack (debian): https://download.elastic.co/elasticsearch/elasticsearch/elasticsearch-1.6.0.deb
- Replace config/elasticsearch.yml by elasticsearch.yml and config/logging.yml by logging.yml from this repo.
- Install plugins with shell script es-plugins.sh under elasticsearch root directory.
- Run ./elasticsearch (with optional heapsize settings: -Xmx2G -Xms2G)
- Go to http://:9200/_plugin/HQ to see cluster status
- Register a repository with a PUT to /_snapshot/<REPO_NAME> with body as in repo_s3_register.json
- Make a POST to /_snapshot/<REPO_NAME>/<SNAPSHOT_NAME> with optional body as in backup_settings.json. Add ?wait_for_completion=true if you want to wait & see backup summary
- If on another machine/cluster register a repository with a PUT to
/_snapshot/<REPO_NAME>
with body as in repo_s3_register.json - View available snapshots by GET to
/_snapshot/<REPO_NAME>/_all
- Make a POST to
/_snapshot/<REPO_NAME>/<SNAPSHOT_NAME>/_restore
with optional body as in restore_settings.json. Add ?wait_for_completion=true if you want to wait & see restore summary - Note that searching will not be available on under-recovery shards.
- Basic queries can be done via GET to
/<INDEX_NAME>/<DOC_TYPE>/_search?q=<FIELD_NAME>:<SEARCH_KEYWORD>
- Complex queries must be done by POST to
/<INDEX_NAME>/<DOC_TYPE>/_search
with a body carrying specifications