Skip to content

Latest commit

 

History

History
99 lines (67 loc) · 3.11 KB

README.md

File metadata and controls

99 lines (67 loc) · 3.11 KB
Raw

Elasticsearch bash utils

This is a collection of useful scripts for working with Elasticsearch on the command line. They aren't the prettiest tools out there, but they provide information-dense and accurate information crucial to managing real live production Elasticsearch instances.

These scripts were originally made for use with the Pelias Geocoder, but are mostly general purpose.

Quickstart

All scripts default to looking at a local cluster, so to quickly get started monitoring a cluster, the following can be used:

git clone https://github.com/geocodeearth/elasticsearch-bash-utils.git
cd elasticsearch-bash-utils
watch bash es_diagnostic.sh

Other scripts

Common config parameters

cluster_url

All scripts take a cluster_url variable. The default is http://localhost:9200. The URL should include the protocol (http/https), but should not include a trailing slash.

Load a snapshot

This script lets you set up a Snapshot Repository pointing to a location in S3 and load a snapshot in a single command.

Example usage:

s3_bucket="your-s3-bucket" \
base_path="the/path/within/your/s3/bucket/where/the/snapshot/repository/lives" \
snapshot_name="the name of the snapshot, or leave blank to use the first snapshot listed" \
cluster_url="http://the-path-to-your-cluster" \
bash load_snapshot.sh

Load a snapshot (without setting up a repository)

The same as above, but skips setting up a repository.

Example usage:

es_repo_name="your-repo-name" \
snapshot_name="the name of the snapshot, or leave blank to use the first snapshot listed" \
cluster_url="http://the-path-to-your-cluster" \
bash load_snapshot_only.sh

Make an alias

This script creates a pelias alias, which can be helpful when working with multiple snapshots. Pelias can be configured to use any index name you choose, but the default is pelias.

Example usage:

cluster_url="http://the-path-to-your-cluster" \
index_name="yourIndex" \
alias_name="pelias" \
bash make_alias.sh

This will make an alias at pelias pointing to yourIndex.

The alias_name parameter defaults to pelias and can be omitted.

Add (or remove) replica shards

This script can add or remove replica shards, useful when setting up a production ready cluster that needs to be resilient against node failure.

Example usage:

cluster_url="http://the-path-to-your-cluster" \
index_name="pelias" \
replica_count="1" \
bash add_replica.sh

The default value for index_name is pelias, and the default replica_count is 1. You can remove all replicas with replica_count="0".

Force merge an index

Force merging indices is a best practice to optimize performance of read-only workloads.

Example usage:

cluster_url="http://the-path-to-your-cluster" \
index_name="pelias" \
num_segments="1" \
bash force_merge.sh

The default value for num_segments is 1. Higher values will complete faster, but 1 offers optimal performance.