Skip to content

whitfin/limber

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Limber

Build Status Crates.io

A simple (but quick) tool for backing up Elasticsearch documents.

Limber offers a very minimal interface to export and import documents from Elasticsearch clusters. Other tools do exist for this purpose, but they're fairly slow and thus unusable on larger clusters. This tool is intended to be as lightweight and fast as possible (given that it's a CLI tool). Usage is meant to align with Unix style pipes to allow piping directly into other clusters, or your own compression algorithms, etc.

The feature set is deliberately small as other tools exist for those who are interested in sugar rather than speed. If you would like to request a new feature, please file an issue and we can discuss inclusion.

Installation

Limber will be available via Crates.io, so you can install it directly with cargo:

$ cargo install limber

Once I become more familiar with cross compilation, I'll try to attach some pre-built binaries to the repository to make it easier for those outside of the Rust ecosystem.

Usage

This section will cover (at a very high level), the most relevant commands offered by Limber. If you want up to date usage, please check the help menu generated via limber -h in your terminal session.

Exporting Documents

To export documents from an Elasticsearch cluster, you can use the export subcommand. This is formed in the following pattern:

$ limber export <source>

The source parameter is required, and much contain the Elasticsearch URL of the cluster you wish to export from. It is assumed that the cluster is reachable from the machine you're running this tool on. If you wish to export from a specific index, you can also include this on the URL in the form of:

http://localhost:9200/my_index
http://localhost:9200/my_first_index,my_second_index

Exported documents are dumped directly to stdout, and progress will be reported to stderr. This allows you to pipe the results into whatever destination you wish in a streaming fashion. As such, invocation of the export command typically looks something like this:

$ limber export http://localhost:9200 | gzip -9 > export.jsonl.gz

The above command will compress all exported data into the export.jsonl.gz file, whilst occasionally reporting progress to your terminal session (as long as you don't also redirect stderr).

There are several options which can be used to customize the export, such as the concurrency factor, batch sizes, document filtering, etc. All of these options can be found via limber export -h.

Importing Documents

The process of importing documents is extremely similar to exporting them, except it operates in reverse. Just like the export command, this is formed in the following pattern:

$ limber import <target>

The target parameter is required, and works in the same way as the source parameter to control which index you place your documents into. If you do not provide an index, the documents will be placed into the same index name as the one they were exported from. Naturally, using multiple index names no longer makes sense when indexing documents.

In similar fashion to exports, documents to import are read from stdin in order to allow you to stream from one location to another. Below is the an example of importing the documents exported in the example further up:

$ gzcat export.jsonl.gz | limber import http://localhost:9200

You might notice that the API here allows you to pipe from one cluster or index directly into another. As an example:

$ limber export http://localhost:9200/my_first_index | \
    limber import http://localhost:9200/my_second_index

The import command also allows for customization of concurrency factor and batch sizes. For all available options, please see limber import -h.