?> Factotum 0.4.0 released with support for constraints | Snowplow
Snowplow Outperforms GA4 in GigaOm Comparative Evaluation.
Read More
  1. Home
  2. Blog
  3. Releases

Factotum 0.4.0 released with support for constraints

Jump to

We’re pleased to announce the 0.4.0 release of Snowplow’s DAG running tool Factotum! This release centers around making DAGs safer to run on distributed clusters by constraining the run to a specific host.

In the rest of this post we will cover:

  1. Constraining job runs
  2. Downloading and running Factotum
  3. Roadmap
  4. Contributing

1. Constraining job runs

This release adds the ability to constrain your DAG’s execution to a single host. This allows for job distribution to many servers while ensuring that the job is only run once and from one place; you can use this for example to distribute a single cron file to multiple boxes, knowing that only one server’s Factotum instance will successfully invoke a given DAG.

To constrain the DAG simply append the following:

./factotum run ./echo.factotum --constraint "host,${your_hostname}" 

The hostname can be:

  • A wildcard * which means that this check always passes
  • The hostname of the server
  • The internal IP of the server

For example running the echo.factfile sample with constraints that all pass:

$ ./factotum run echo.factfile --constraint "host,*" $ ./factotum run echo.factfile --constraint "host,Joshuas-iMac.local" $ ./factotum run echo.factfile --constraint "host," # Ethernet $ ./factotum run echo.factfile --constraint "host," # WiFi 

However, if I change these to non-valid values:

$ ./factotum run echo.factfile --constraint "host,hostname" $ Warn: the specifed host constraint "hostname" did not match, no tasks have been executed. Reason: could not find any IPv4 addresses for the supplied hostname 

This means that we failed to lookup any addresses associated with the hostname given: we have nothing to compare against our local interfaces and as such we cannot obtain a match.

If I were to disconnect my WiFi network interface:

$ ./factotum run echo.factfile --constraint "host," $ Warn: the specifed host constraint "" did not match, no tasks have been executed. Reason: failed to match any of the interface addresses to the found host addresses 

This means that we were able to get addresses from the local interfaces as well as from the supplied IP address, but a match could not be found within both lists.

Note: A failure to match the host constraint will result in an exit code of 0. This is because a non-matching host is considered a no-operation (“noop”).

2. Downloading and running Factotum

Factotum is available for macOS and Linux (x86_64).

The following series of commands will download the 0.4.0 release, unzip it in your current working directory and download a sample job for you to run.

If you’re running Linux:

$ wget https://bintray.com/artifact/download/snowplow/snowplow-generic/factotum_0.4.0_linux_x86_64.zip $ unzip factotum_0.4.0_linux_x86_64.zip $ wget https://raw.githubusercontent.com/snowplow/factotum/master/samples/echo.factfile 

If you’re running macOS:

$ wget https://bintray.com/artifact/download/snowplow/snowplow-generic/factotum_0.4.0_darwin_x86_64.zip $ unzip factotum_0.4.0_darwin_x86_64.zip $ wget https://raw.githubusercontent.com/snowplow/factotum/master/samples/echo.factfile 

You can then run Factotum in the following way:

$ ./factotum run ./echo.factfile 

3. Roadmap for Factotum

We’re taking an iterative approach with Factotum – today Factotum won’t give you an entire stack for monitoring, scheduling and running data pipelines, but we plan on growing it into a set of tools that will.

Coming up soon, we are looking forward to open-sourcing a prototype job server for Factotum, which allows you to submit jobs for Factotum to run over HTTP.

If you have specific features to suggest, please add a ticket to the GitHub repo.

4. Contributing

Factotum is completely open source – and has been from the start! If you’d like to get involved, or just try your hand at Rust, please check out the repository.

More about
the author

Joshua Beemster
Joshua Beemster

Josh is a Head of Engineering and enjoys mapping out how to manage infrastructure across multiple clouds using a combination of Terraform, K8s and Helm. He works mainly with the teams responsible for infrastructure and the core data processing teams at Snowplow.

View author

Ready to start creating rich, first-party data?

Image of the Snowplow app UI