This looks like a great way to learn Ansible too. Instead of learning alongside random examples, you can setup your server and see how it would look like in Ansible.
Indeed when using single-shot, unless you're using the --remote modes (in which case, the harvest is pulled down to a machine-generated path locally), indeed you need to supply the path to the harvest so that the 'manifest' part under the hood, knows what to use.
(By contrast, if you are using just the 'enroll harvest' command by itself, and omit the --out option, it will by default store the harvest in a random directory in ~/.cache/enroll/harvest/xxxxxxx)
This is a great idea. I have done this manually, and it was a lot of work.
Even with a tool, people will still have to understand the output, enough that they can spot situations like "this part doesn't make sense at all", "that bit isn't static", "holy crud, there's an unsecured secret", "this part suggests a dependency on this other server we didn't know was involved, and which the tool doesn't investigate".
I agree! It's always a 'best effort' tool. There's going to be corner cases where something that might end up in the 'logrotate' role could arguably be better placed in a more specific app's role.
It does an okay job at this sort of thing, but definitely human eyes are needed :)
Bravo, I will play with it. I haven't played with ansible till now but I know that its related to automation.
If something can make ansible easier for me to try out like this tool while being pragmatic, I will give this a try someday thank you!
How accurate does this tool end up becoming though? Like can I just run some bunch of commands to setup a server and then use this with ansible?
Would this end up being a good use for it or would I just re-invent something similar to cloud-init on wrong abstraction. (On all fairness, one thing troubling me about cloud-init is that I would need to probably have a list of all commands that I want to run and all changes which sometimes if history command might have some issues or you do end up writing files etc. ends up being a little messy)
I haven't played that much with both cloud-init and ansible either but I am super interested to know more about enroll and others as well as I found it really cool!
Great questions! OP here, let me answer them below:
> How accurate does this tool end up becoming though? Like can I just run some bunch of commands to setup a server and then use this with ansible?
Yes, exactly: let's say you provision a VPS and then install some stuff on it, configure some configs, create a crontab, create a user account. Running 'enroll harvest' on it will detect all of that, and 'enroll manifest' will then convert that 'harvest' into Ansible roles/playbooks.
> Would this end up being a good use for it or would I just re-invent something similar to cloud-init on wrong abstraction. (On all fairness, one thing troubling me about cloud-init is that I would need to probably have a list of all commands that I want to run and all changes which sometimes if history command might have some issues or you do end up writing files etc. ends up being a little messy)
Yeah, your instinct is right on the latter point. Ansible and Cloud-init are similar in that they are both 'declarative' systems to say what should exist on the machine. Ansible has some advantages in that it compares with the current setup to see if it needs to change anything. Cloud-init (in my experience) is a bit more crude: 'just run this stuff the first time the machine is booted'.
I'm sure there are different ways of using it, but in my experience, cloud-init is really designed to 'run once' (first time setup). For example, if you provision a machine with Terraform or OpenTofu, and you pass in cloud-init, then later if you change the cloud-init data, it wants to destroy the machine and rebuild it (unless you configure it explicitly not to do that, by which you have to tell it to 'ignore' changes to the cloud-init).
Whereas with Ansible, you're at least equipped with a solid foundation that you can extend over time - you'll no doubt eventually need to make changes to your server post the initial setup.
If you're new to Ansible, Enroll will be a quick way to get up and running with working Ansible configuration and you can adapt it from there as you learn it.
Admittedly, to satisfy a lot of corner cases (or support different Linux distros), the Ansible code that Enroll generates is a bit complex to read compared to a 'bespoke' home-grown playbook, on the other hand, it's perfectly correct code and you'd be immediately exposed to good practice syntax.
Does the playbook generation have support for some totally custom/one-off application? (Eg, not just system/well-known stuff). If so, that would be insane!
It does! There are several sort of 'catch-alls' in place:
1) stuff in /etc that doesn't belong to any obvious package, ends up in an 'etc_custom' role
2) stuff in /usr/local ends up in a 'usr_local_custom' role
3) Anything you include with --include will end up in a special 'extra_paths' role.
Here's a demo (which is good, helped me spot a small bug, the role is included twice in the playbook :) I'll get that fixed!) https://asciinema.org/a/765385
This is a fantastic idea. I can imagine using this to pull in any manual changes I might have made to the server because I’m not the most disciplined person.
Haha, same! I ran it on a server I've been shepherding along since 2008 and wow, it was insightful, there were even cron jobs it found that I had forgotten about :)
If you are using a Debian-like or Fedora-like workstation, it's also really useful to 'ansibilize' your desktop OS in case you need to reinstall :)
Indeed! I'm showing my age, but I do remember using this with Puppet and it was one of my inspirations :D (no commits in nearly 13 years, ouch) https://github.com/devstructure/blueprint
Yes! I always thought that was a very clever project, and was sad when it ceased development. Very excited to try this out, and glad to have stayed on Debian all these years.
I have quite a few machines that were constructed using Ansible ... When I get a chance, I'll reverse then and compare the results to the IaC that created them
Wonderful! I wish that tool was existed a few years ago, when I had no experience with Ansible. Anyway, will try it and compare outcome of Enroll with my current playbooks
Could it also detect changed package files; if there are per-package-file checksums like with `debsums` and `rpm -V`?
Does it check extended filesystem labels with e.g. getfacl for SELinux support?
I've also done this more than a few times and not written a tool.
At least once I've scripted better then regex to convert a configuration file to a Jinja2 templated configuration file (from the current package's default commented config file with the latest options). And then the need is to diff: non-executable and executable guidelines, the package default config (on each platform), and our config.
Sometimes it's better not to re-specify a default config param and value, but only if the defaults are sane on every platform. Cipher lists for example.
P2V (physical to virtual) workflows don't result in auditable system policy like this.
Most of the OS and Userspace packages backed up in full system images (as with typical P2V workflows) are exploitably out of date in weeks or months.
To do immutable upgrades with rollback,
Rpm-ostree distros install the RPM packages atop the latest signed immutable rootfs image, and then layer /etc on top (and mounts in /var which hosts flatpaks and /var/home). It keeps a list of packages to reinstall and it does a smart merge of /etc. Unfortunately etckeeper (which auto-git-commits /etc before and after package upgrades) doesn't yet work with rpm-ostree distros.
Ansible does not yet work with rpm-ostree distros. IIRC the primary challenge is that ansible wants to run each `dnf install` individually and that takes forever with rpm-ostree. It is or is not the same to install one long list of packages or to install multiple groups of packages in the same sequence. It should be equivalent if the package install and post-install scripts are idempotent, but is not equivalent if e.g. `useradd` is called multiply without an explicit UID in package scripts which run as root too.
I wrote a PR to get structured output (JSON) from `dnf history`, but it was for dnf4.
> upgrading the layered firefox RPM without a reboot requires -A/--apply-live (which runs twice) and upgrading the firefox flatpak doesn't require a reboot, but SELinux policies don't apply to flatpaks which run unconfined FWIU.
Does it log a list of running processes and their contexts; with `ps -Z`?
There are also VM-level diff'ing utilities for forensic-level differencing.
It also tries to ignore packages that came with the distro automatically, e.g focusing on stuff that was explicitly installed (based on 'apt-mark showmanual' for Debian, and 'dnf -q repoquery --userinstalled' (and related commands, like dnf -q history userinstalled) for RH-like)
> Does it check extended filesystem labels with e.g. getfacl for SELinux support?
Not yet, but that's interesting, I'll look into it.
> At least once I've scripted better then regex to convert a configuration file to a Jinja2 templated configuration file (from the current package's default commented config file with the latest options).
Yep, that was the inspiration for my companion tool https://git.mig5.net/mig5/jinjaturtle (which enroll will automatically try and use if it finds it on the $PATH - if it can't find it, it will just use 'copy' mode for Ansible tasks, and the original files).
Note that running the `enroll manifest` command against multiple separate 'harvests' (e.g harvested from separate machines) but storing it in the same common manifest location, will 'merge' the Ansible manifests. Thereby 'growing' the Ansible manifest as needed. But each host 'feature flips' on/off which files/templates should be deployed on it, based on what was 'harvested' from that host.
> Does it log a list of running processes and their contexts; with `ps -Z`?
Very cool! Managing ones boxes as cattle and not pets almost always seems like a better idea in retrospect but historically it is easier said than done. Moreover, I like the idea of being able to diff a box's actual state from a current Ansible system to verify that it actually is as configured for further parity between deployed/planned.
Definitely! It's all too easy to make a direct change and later forget to 'fold it in' to Ansible and run a playbook. My hope is that `enroll diff` serves as a good reminder if nothing else.
I'm pondering adding some sort of `--enforce` argument to make it re-apply a 'golden' harvest state if you really want to be strictly against drift. For now, it's notifications only though.
Awesome stuff!
https://asciinema.org/a/765293
Enroll will automatically make use of jinjaturtle if it's on the $PATH, to generate templates.
I just saved the state of my WSL2 instance, pushed it to github. Amazingly simple.
FWIW, I was required to add the --harvest, which your quick start seems to be missing?
ie I used:
uvx enroll single-shot --harvest ./harvest --out ./ansible
Indeed when using single-shot, unless you're using the --remote modes (in which case, the harvest is pulled down to a machine-generated path locally), indeed you need to supply the path to the harvest so that the 'manifest' part under the hood, knows what to use.
(By contrast, if you are using just the 'enroll harvest' command by itself, and omit the --out option, it will by default store the harvest in a random directory in ~/.cache/enroll/harvest/xxxxxxx)
Thanks for trying it out!
Even with a tool, people will still have to understand the output, enough that they can spot situations like "this part doesn't make sense at all", "that bit isn't static", "holy crud, there's an unsecured secret", "this part suggests a dependency on this other server we didn't know was involved, and which the tool doesn't investigate".
It does an okay job at this sort of thing, but definitely human eyes are needed :)
If something can make ansible easier for me to try out like this tool while being pragmatic, I will give this a try someday thank you!
How accurate does this tool end up becoming though? Like can I just run some bunch of commands to setup a server and then use this with ansible?
Would this end up being a good use for it or would I just re-invent something similar to cloud-init on wrong abstraction. (On all fairness, one thing troubling me about cloud-init is that I would need to probably have a list of all commands that I want to run and all changes which sometimes if history command might have some issues or you do end up writing files etc. ends up being a little messy)
I haven't played that much with both cloud-init and ansible either but I am super interested to know more about enroll and others as well as I found it really cool!
> How accurate does this tool end up becoming though? Like can I just run some bunch of commands to setup a server and then use this with ansible?
Yes, exactly: let's say you provision a VPS and then install some stuff on it, configure some configs, create a crontab, create a user account. Running 'enroll harvest' on it will detect all of that, and 'enroll manifest' will then convert that 'harvest' into Ansible roles/playbooks.
> Would this end up being a good use for it or would I just re-invent something similar to cloud-init on wrong abstraction. (On all fairness, one thing troubling me about cloud-init is that I would need to probably have a list of all commands that I want to run and all changes which sometimes if history command might have some issues or you do end up writing files etc. ends up being a little messy)
Yeah, your instinct is right on the latter point. Ansible and Cloud-init are similar in that they are both 'declarative' systems to say what should exist on the machine. Ansible has some advantages in that it compares with the current setup to see if it needs to change anything. Cloud-init (in my experience) is a bit more crude: 'just run this stuff the first time the machine is booted'.
I'm sure there are different ways of using it, but in my experience, cloud-init is really designed to 'run once' (first time setup). For example, if you provision a machine with Terraform or OpenTofu, and you pass in cloud-init, then later if you change the cloud-init data, it wants to destroy the machine and rebuild it (unless you configure it explicitly not to do that, by which you have to tell it to 'ignore' changes to the cloud-init).
Whereas with Ansible, you're at least equipped with a solid foundation that you can extend over time - you'll no doubt eventually need to make changes to your server post the initial setup.
If you're new to Ansible, Enroll will be a quick way to get up and running with working Ansible configuration and you can adapt it from there as you learn it.
Admittedly, to satisfy a lot of corner cases (or support different Linux distros), the Ansible code that Enroll generates is a bit complex to read compared to a 'bespoke' home-grown playbook, on the other hand, it's perfectly correct code and you'd be immediately exposed to good practice syntax.
Let me know if you get to try it!
1) stuff in /etc that doesn't belong to any obvious package, ends up in an 'etc_custom' role
2) stuff in /usr/local ends up in a 'usr_local_custom' role
3) Anything you include with --include will end up in a special 'extra_paths' role.
Here's a demo (which is good, helped me spot a small bug, the role is included twice in the playbook :) I'll get that fixed!) https://asciinema.org/a/765385
Thanks for your interest!
If you are using a Debian-like or Fedora-like workstation, it's also really useful to 'ansibilize' your desktop OS in case you need to reinstall :)
Does it check extended filesystem labels with e.g. getfacl for SELinux support?
I've also done this more than a few times and not written a tool.
At least once I've scripted better then regex to convert a configuration file to a Jinja2 templated configuration file (from the current package's default commented config file with the latest options). And then the need is to diff: non-executable and executable guidelines, the package default config (on each platform), and our config.
Sometimes it's better not to re-specify a default config param and value, but only if the defaults are sane on every platform. Cipher lists for example.
P2V (physical to virtual) workflows don't result in auditable system policy like this.
Most of the OS and Userspace packages backed up in full system images (as with typical P2V workflows) are exploitably out of date in weeks or months.
To do immutable upgrades with rollback, Rpm-ostree distros install the RPM packages atop the latest signed immutable rootfs image, and then layer /etc on top (and mounts in /var which hosts flatpaks and /var/home). It keeps a list of packages to reinstall and it does a smart merge of /etc. Unfortunately etckeeper (which auto-git-commits /etc before and after package upgrades) doesn't yet work with rpm-ostree distros.
Ansible does not yet work with rpm-ostree distros. IIRC the primary challenge is that ansible wants to run each `dnf install` individually and that takes forever with rpm-ostree. It is or is not the same to install one long list of packages or to install multiple groups of packages in the same sequence. It should be equivalent if the package install and post-install scripts are idempotent, but is not equivalent if e.g. `useradd` is called multiply without an explicit UID in package scripts which run as root too.
I wrote a PR to get structured output (JSON) from `dnf history`, but it was for dnf4.
From https://news.ycombinator.com/item?id=43617363 :
> upgrading the layered firefox RPM without a reboot requires -A/--apply-live (which runs twice) and upgrading the firefox flatpak doesn't require a reboot, but SELinux policies don't apply to flatpaks which run unconfined FWIU.
Does it log a list of running processes and their contexts; with `ps -Z`?
There are also VM-level diff'ing utilities for forensic-level differencing.
> Could it also detect changed package files; if there are per-package-file checksums like with debsums and `rpm -V`?
Yes, that's exactly what it does. See https://git.mig5.net/mig5/enroll/src/branch/main/enroll/plat... and https://git.mig5.net/mig5/enroll/src/branch/main/enroll/rpm....
It also tries to ignore packages that came with the distro automatically, e.g focusing on stuff that was explicitly installed (based on 'apt-mark showmanual' for Debian, and 'dnf -q repoquery --userinstalled' (and related commands, like dnf -q history userinstalled) for RH-like)
> Does it check extended filesystem labels with e.g. getfacl for SELinux support?
Not yet, but that's interesting, I'll look into it.
> At least once I've scripted better then regex to convert a configuration file to a Jinja2 templated configuration file (from the current package's default commented config file with the latest options).
Yep, that was the inspiration for my companion tool https://git.mig5.net/mig5/jinjaturtle (which enroll will automatically try and use if it finds it on the $PATH - if it can't find it, it will just use 'copy' mode for Ansible tasks, and the original files).
Note that running the `enroll manifest` command against multiple separate 'harvests' (e.g harvested from separate machines) but storing it in the same common manifest location, will 'merge' the Ansible manifests. Thereby 'growing' the Ansible manifest as needed. But each host 'feature flips' on/off which files/templates should be deployed on it, based on what was 'harvested' from that host.
> Does it log a list of running processes and their contexts; with `ps -Z`?
It doesn't use ps, but it examines systemctl to get a list of running services and also timers. Have a look at https://git.mig5.net/mig5/enroll/src/branch/main/enroll/syst...
Thanks for the other ideas! I'll look into them.
I'm pondering adding some sort of `--enforce` argument to make it re-apply a 'golden' harvest state if you really want to be strictly against drift. For now, it's notifications only though.
- Sent from my NixOS daily-driver, which only cost me a small number of grey hairs