5 stable releases
|1.1.1||Jun 15, 2022|
|1.1.0||Apr 5, 2022|
|1.0.2||Jul 22, 2019|
|1.0.1||Jun 7, 2019|
|1.0.0||May 24, 2019|
#722 in Command line utilities
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.
Limber will be available via Crates.io,
so you can install it directly with
$ 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.
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
limber -h in your terminal session.
To export documents from an Elasticsearch cluster, you can use the
subcommand. This is formed in the following pattern:
$ limber export <source>
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
Exported documents are dumped directly to
stdout, and progress will be
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
file, whilst occasionally reporting progress to your terminal session (as
long as you don't also redirect
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.
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>
target parameter is required, and works in the same way as the
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
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.