This project provides a convenient way to populate values from Consul
into the file system using the consul-template
daemon.
The daemon consul-template
queries a Consul or Vault
cluster and updates any number of specified templates on the file system. As an
added bonus, it can optionally run arbitrary commands when the update process
completes. Please see the examples folder for some scenarios where
this functionality might prove useful.
The documentation in this README corresponds to the master branch of Consul Template. It may contain unreleased features or different APIs than the most recently released version.
Please see the Git tag that corresponds to your version of Consul Template for the proper documentation.
- Community Support
- Installation
- Quick Example
- Learn Guides
- Configuration
- Command Line Flags
- Configuration File
- Reload Configuration and Templates
- Templating Language
- API Functions
- Scratch
- Helper Functions
- Math Functions
- Observability
- Logging
- Modes
- Once Mode
- De-Duplication Mode
- Exec Mode
- Plugins
- Caveats
- Docker Image Use
- Dots in Service Names
- Termination on Error
- Commands
- Multi-phase Execution
- Debugging
- FAQ
- Contributing
If you have questions about how consul-template works, its capabilities or anything other than a bug or feature request (use github's issue tracker for those), please see our community support resources.
Community portal: https://discuss.hashicorp.com/tags/c/consul/29/consul-template
Other resources: https://www.consul.io/community.html
Additionally, for issues and pull requests, we'll be using the 👍 reactions as a rough voting system to help gauge community priorities. So please add 👍 to any issue or pull request you'd like to see worked on. Thanks.
-
Download a pre-compiled, released version from the Consul Template releases page.
-
Extract the binary using
unzip
ortar
. -
Move the binary into
$PATH
.
To compile from source, please see the instructions in the contributing section.
This short example assumes Consul is installed locally.
- Start a Consul cluster in dev mode:
$ consul agent -dev
- Author a template
in.tpl
to query the kv store:
{{ key "foo" }}
- Start Consul Template:
$ consul-template -template "in.tpl:out.txt" -once
- Write data to the key in Consul:
$ consul kv put foo bar
- Observe Consul Template has written the file
out.txt
:
$ cat out.txt
bar
For more examples and use cases, please see the examples folder in this repository.
In addition to these examples, HashiCorp has published guides and official documentation to help walk through a few common use cases for Consul Template.
Configuration documentation has been moved to docs/configuration.md.
While there are multiple ways to run Consul Template, the most common pattern is to run Consul Template as a system service. When Consul Template first starts, it reads any configuration files and templates from disk and loads them into memory. From that point forward, changes to the files on disk do not propagate to running process without a reload.
The reason for this behavior is simple and aligns with other tools like haproxy.
A user may want to perform pre-flight validation checks on the configuration or
templates before loading them into the process. Additionally, a user may want to
update configuration and templates simultaneously. Having Consul Template
automatically watch and reload those files on changes is both operationally
dangerous and against some of the paradigms of modern infrastructure. Instead,
Consul Template listens for the SIGHUP
syscall to trigger a configuration
reload. If you update configuration or templates, simply send HUP
to the
running Consul Template process and Consul Template will reload all the
configurations and templates from disk.
Templating Language documentation has been moved to docs/templating-language.md.
The Alpine Docker image is configured to support an external volume to render shared templates to. If mounted you will need to make sure that the consul-template user in the docker image has write permissions to the directory. Also if you build your own image using these you need to be sure you have the permissions correct.
The consul-template user in docker has a UID of 100 and a GID of 1000.
This effects the in image directories /consul-template/config, used to add configuration when using this as a parent image, and /consul-template/data, exported as a VOLUME as a location to render shared results.
Previously the image initially ran as root in order to ensure the permissions allowed it. But this ran against docker best practices and security policies.
If you build your own image based on ours you can override these values with
--build-arg
parameters.
Using dots .
in service names will conflict with the use of dots for TAG
delineation in the
template. Dots already interfere with using
DNS
for service names, so we recommend avoiding dots wherever possible.
By default Consul Template is highly fault-tolerant. If Consul is unreachable or
a template changes, Consul Template will happily continue running. The only
exception to this rule is if the optional command
exits non-zero. In this
case, Consul Template will also exit non-zero. The reason for this decision is
so the user can easily configure something like Upstart or God to manage Consul
Template as a service.
If you want Consul Template to continue watching for changes, even if the
optional command argument fails, you can append || true
to your command. Note
that ||
is a "shell-ism", not a built-in function. You will also need to run
your command under a shell:
$ consul-template \
-template "in.ctmpl:out.file:/bin/bash -c 'service nginx restart || true'"
In this example, even if the Nginx restart command returns non-zero, the overall
function will still return an OK exit code; Consul Template will continue to run
as a service. Additionally, if you have complex logic for restarting your
service, you can intelligently choose when you want Consul Template to exit and
when you want it to continue to watch for changes. For these types of complex
scripts, we recommend using a custom sh or bash script instead of putting the
logic directly in the consul-template
command or configuration file.
The current processes environment is used when executing commands with the following additional environment variables:
CONSUL_HTTP_ADDR
CONSUL_HTTP_TOKEN
CONSUL_HTTP_AUTH
CONSUL_HTTP_SSL
CONSUL_HTTP_SSL_VERIFY
These environment variables are exported with their current values when the
command executes. Other Consul tooling reads these environment variables,
providing smooth integration with other Consul tools (like consul maint
or
consul lock
). Additionally, exposing these environment variables gives power
users the ability to further customize their command script.
The command configured for running on template rendering must take one of two forms.
The first is as a single command without spaces in its name and no arguments. This form of command will be called directly by consul-template and is good for any situation. The command can be a shell script or an executable, anything called via a single word, and must be either on the runtime search PATH or the absolute path to the executable. The single word limination is necessary to eliminate any need for parsing the command line. For example..
command = "/opt/foo"
or, if on PATH, command = "foo"
The second form is as a multi-word command, a command with arguments or a more
complex shell command. This form requires a shell named sh
be on the
executable search path (eg. PATH on *nix). This is the standard on all *nix
systems and should work out of the box on those systems. This won't work on,
for example, Docker images with only the executable and not a minimal system
like Alpine. Using this form you can join multiple commands with logical
operators, &&
and ||
, use pipelines with |
, conditionals, etc. Note that
the shell sh
is normally /bin/sh
on *nix systems and is either a POSIX
shell or a shell run in POSIX compatible mode, so it is best to stick to POSIX
shell syntax in this command. For example..
command = "/opt/foo && /opt/bar"
command = "if /opt/foo ; then /opt/bar ; fi"
Using this method you can run as many shell commands as you need with whatever logic you need. Though it is suggested that if it gets too long you might want to wrap it in a shell script, deploy and run that.
Consul Template does an n-pass evaluation of templates, accumulating dependencies on each pass. This is required due to nested dependencies, such as:
{{ range services }}
{{ range service .Name }}
{{ .Address }}
{{ end }}{{ end }}
During the first pass, Consul Template does not know any of the services in Consul, so it has to perform a query. When those results are returned, the inner-loop is then evaluated with that result, potentially creating more queries and watches.
Because of this implementation, template functions need a default value that is
an acceptable parameter to a range
function (or similar), but does not
actually execute the inner loop (which would cause a panic). This is important
to mention because complex templates must account for the "empty" case. For
example, the following will not work:
{{ with index (service "foo") 0 }}
# ...
{{ end }}
This will raise an error like:
<index $services 0>: error calling index: index out of range: 0
That is because, during the first evaluation of the template, the service
key is returning an empty slice. You can account for this in your template like
so:
{{ with service "foo" }}
{{ with index . 0 }}
{{ .Node }}{{ end }}{{ end }}
This will still add the dependency to the list of watches, but will not evaluate the inner-if, avoiding the out-of-index error.
Consul Template can print verbose debugging output. To set the log level for
Consul Template, use the -log-level
flag:
$ consul-template -log-level info ...
<timestamp> [INFO] (cli) received redis from Watcher
<timestamp> [INFO] (cli) invoking Runner
# ...
You can also specify the level as debug:
$ consul-template -log-level debug ...
<timestamp> [DEBUG] (cli) creating Runner
<timestamp> [DEBUG] (cli) creating Consul API client
<timestamp> [DEBUG] (cli) creating Watcher
<timestamp> [DEBUG] (cli) looping for data
<timestamp> [DEBUG] (watcher) starting watch
<timestamp> [DEBUG] (watcher) all pollers have started, waiting for finish
<timestamp> [DEBUG] (redis) starting poll
<timestamp> [DEBUG] (service redis) querying Consul with &{...}
<timestamp> [DEBUG] (service redis) Consul returned 2 services
<timestamp> [DEBUG] (redis) writing data to channel
<timestamp> [DEBUG] (redis) starting poll
<timestamp> [INFO] (cli) received redis from Watcher
<timestamp> [INFO] (cli) invoking Runner
<timestamp> [DEBUG] (service redis) querying Consul with &{...}
# ...
Q: How is this different than confd?
A: The answer is simple: Service Discovery as a first class citizen. You are also encouraged to read this Pull Request on the project for more background information. We think confd is a great project, but Consul Template fills a missing gap. Additionally, Consul Template has first class integration with Vault, making it easy to incorporate secret material like database credentials or API tokens into configuration files.
Q: How is this different than Puppet/Chef/Ansible/Salt?
A: Configuration management tools are designed to be used in unison with Consul Template. Instead of rendering a stale configuration file, use your configuration management software to render a dynamic template that will be populated by Consul.
To build and install Consul-Template locally, you will need to install Go.
Clone the repository:
$ git clone https://github.com/hashicorp/consul-template.git
To compile the consul-template
binary for your local machine:
$ make dev
This will compile the consul-template
binary into bin/consul-template
as
well as your $GOPATH
and run the test suite.
If you want to compile a specific binary, set XC_OS
and XC_ARCH
or run the
following to generate all binaries:
$ make build
If you want to run the tests, first install consul and vault locally, then:
$ make test
Or to run a specific test in the suite:
go test ./... -run SomeTestFunction_name