Best practice around putting rails database info into git

rubyonrails-talk
1

This maybe too vague a question, but I wonder if there's some wisdom out there for me to tap...

Sometimes when I get ready to commit, git says that db/schema.rb has changed. This is weird, and I think it's because I did a rake db:migrate.

So my question is, what's the best practice relating to putting bits of the database (schema.rb and the databases themselves) into sourc control. And what's the best practice relating to running a rake db:migrate after a git checkout of a new branch?

Anyone?

- Pito

2

  1. I usually put my db/schema.rb to .gitignore. There’s some rake task to create your database not from migrations, but from schema.rb. But I didn’t see anybody alive doing this. So, you don’t really need this file in your repo.

  2. You run your new migrations as soon as you get them. So, after pulling new changes from remote branch

3

Vladimir Rybas wrote:

1. I usually put my db/schema.rb to .gitignore. There's some rake task to create your database not from migrations, but from schema.rb. But I didn't see anybody alive doing this. So, you don't really need this file in your repo.

2. You run your new migrations as soon as you get them. So, after pulling new changes from remote branch

Thanks!!

-- Pito

4

Pito Salas wrote:

This maybe too vague a question, but I wonder if there's some wisdom out there for me to tap...

Sometimes when I get ready to commit, git says that db/schema.rb has changed. This is weird, and I think it's because I did a rake db:migrate.

So my question is, what's the best practice relating to putting bits of the database (schema.rb and the databases themselves) into sourc control. And what's the best practice relating to running a rake db:migrate after a git checkout of a new branch?

Beyond the issues with db/schema.rb, I think it's also good practice to not version control config/database.yml. You may be doing this already.

The reason for this is that database configurations for other member of the team may be different, and you may not want to expose the production database configuration to then entire team.

I do the following:

1. Rename a basic config/database.yml file to config/example_database.yml. 2. Add config/example_database.yml to Git 3. Ignore config/database.yml in .gitignore 4. Commit these changes 5. Copy config/example_database.yml to config/database.yml

When other team members clone the repository they will also need to copy the example_database.yml file and configure it for their own system. This allow more freedom to other developers on the team to setup their system the way they see fit.

For example: I may prefer to use the MySQL root user with no password set on my development database, while other may prefer (or need) their development database protected by using a different user with their own password. Or, maybe I might prefer using SQLite in development while others on the team prefer developing using MySQL.

Keeping this configuration out of the versioning system makes these kinds of preference easier to manage. This technique also helps in cases where not all team member are allowed to know the production database's password.

5

Vladimir Rybas wrote:

1. I usually put my db/schema.rb to .gitignore. There's some rake task to create your database not from migrations, but from schema.rb. But I didn't see anybody alive doing this.

I do it.

So, you don't really need this file in your repo.

Completely wrong. You should always create your new installations' databases with rake db:schema:load , *never* by running migrations from zero. Migrations are only for changing *existing* installations, not for creating new ones from scratch.

Therefore, you need db/schema.rb in version control. There's even a comment in that file that says as much.

config/database.yml shouldn't be in version control because it contains passwords, but I like to put a sample version of that file (without passwords) in the repository as an example of how it should be set up.

2. You run your new migrations as soon as you get them. So, after pulling new changes from remote branch

Right.

Best,

6

Pito Salas wrote:

This maybe too vague a question, but I wonder if there's some wisdom out there for me to tap...

Sometimes when I get ready to commit, git says that db/schema.rb has changed. This is weird, and I think it's because I did a rake db:migrate.

Yes, that would probably be why. Why do you think this is "weird"? It's normal behavior -- schema.rb contains what Rails thinks your database schema is.

So my question is, what's the best practice relating to putting bits of the database (schema.rb and the databases themselves)

schema.rb is not a "bit of the database".

into sourc control. And what's the best practice relating to running a rake db:migrate after a git checkout of a new branch?

To summarize my earlier posts: schema.rb definitely goes into version control. I don't bother putting SQLite databases themselves in the repository, though (not that I use SQLite much).

Anyone?

- Pito

Best,

7

Pito,

I'm going to leave the entire db version control question for someone else, but to answer your question about the schema.rb, if you run a migration, it will change the structure of the database, which is what schema.rb keeps track of. Rails will rewrite that file every for every migration ran to reflect the current structure of the db ( structure as in tables, column names, indexes, etc... Not data the records stored in the db ). This file should be checked into git, and like any file that changes under version control, git will let you know about the changes and track them.

As your app grows, your migration files will be more prone to hving problems and shouldn't be used to create / setup the database. The schema file is much better at recreating the db structure when starting fresh ( like if you have to scale up your app and create a new db on another server ). There are rake tasks for this, such as rake db:reset ( caution, that one will wipe out all data in db), so read up on them and they can save you many headaches as you have to move your databases around.

-Matt

8

Matt,

Thanks for a great summary; very helpful, and thanks to the rest of the list for always providing insightful comments and solutions to problems. Great!!

- Pito

Matt wrote:

9

Matt, Pito, Marnen, and anyone else,

1. The opinion on whether db/schema.rb goes into the source repository has changed over time. It is an opinion. (I'll give you some of mine.)

2. When properly written, migration files are not "prone to having problems" and certainly do not have to get worse over time. The advise here is to define at least a minimal version of your AR model class nested within the migration class if you need to do *any* manipulation at the model level.

3. I think that db/schema.rb does *NOT* belong in the repository. This is particularly true in a multi-developer environment where migrations are being initially created on different branches. The database itself isn't going into the repository after all. If you need the file, run a rake db:schema:dump or just run migrations. If you think about why the migration numbering (file naming) was changed from sequence number to timestamp, you'll realize that the practice of such "interleaved" migrations was a much bigger pain-point than what to do about db/schema.rb.

4. Unless you're initializing (inserting) data via migration, running all the migrations is really not much different than doing a db:schema:load because all the migrations are operating on empty tables. (Besides, if you have to "scale up your app", you probably aren't adding a new empty database, but creating a master-slave or sharding for performance.)

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

10

Rob Biedenharn wrote: [...]

Matt, Pito, Marnen, and anyone else,

1. The opinion on whether db/schema.rb goes into the source repository has changed over time.

No. I've used Rails since 1.2.6. Every version has put a comment in the schema.rb file that recommends putting it into version control.

It is an opinion. (I'll give you some of mine.)

Opinion is not an excuse for advocating dangerous practices.

2. When properly written, migration files are not "prone to having problems" and certainly do not have to get worse over time. The advise here is to define at least a minimal version of your AR model class nested within the migration class if you need to do *any* manipulation at the model level.

But you should never be running old migrations in the first place. If you need version 1000 of the schema for a new installation, then don't start at zero and run 1000 migrations -- just do rake db:schema:load and have done with it. This is the core team's recommendation, and I think it's a good one.

3. I think that db/schema.rb does *NOT* belong in the repository.

Why not? Without it, you can't load the schema with Rake.

This is particularly true in a multi-developer environment where migrations are being initially created on different branches.

I don't see how this makes a difference. Your VCS should be able to merge the schema.rb files. If not, get a better VCS.

The database itself isn't going into the repository after all.

That's a red herring.

If you need the file, run a rake db:schema:dump or just run migrations.

No! You need the file so that rake db:schema:load is possible. You've got it completely backwards.

If you think about why the migration numbering (file naming) was changed from sequence number to timestamp, you'll realize that the practice of such "interleaved" migrations was a much bigger pain-point than what to do about db/schema.rb.

I don't really understand what you're getting at here.

4. Unless you're initializing (inserting) data via migration, running all the migrations is really not much different than doing a db:schema:load because all the migrations are operating on empty tables.

How can you say this with a straight face? There is no reason *at all* to run lots of migrations rather than doing a simple schema load.

(Besides, if you have to "scale up your app", you probably aren't adding a new empty database, but creating a master-slave or sharding for performance.)

Another red herring.

Rob, I know you know a lot about the Rails framework, but your advice here will make dealing with databases far more difficult than it needs to be.

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

Best,

11

Rob Biedenharn wrote: [...]

Matt, Pito, Marnen, and anyone else,

1. The opinion on whether db/schema.rb goes into the source repository has changed over time.

No. I've used Rails since 1.2.6. Every version has put a comment in the schema.rb file that recommends putting it into version control.

I've been using Rails since 0.13 and I'll restate that this opinion has changed. I can't find it at the moment, but there WAS a version that specifically said in that comment to NOT put the file into source control. At least in 1.2.2, there is no comment either way.

It is an opinion. (I'll give you some of mine.)

Opinion is not an excuse for advocating dangerous practices.

"Dangerous"? Well, that's certainly an opinion. :wink:

2. When properly written, migration files are not "prone to having problems" and certainly do not have to get worse over time. The advise here is to define at least a minimal version of your AR model class nested within the migration class if you need to do *any* manipulation at the model level.

But you should never be running old migrations in the first place. If you need version 1000 of the schema for a new installation, then don't start at zero and run 1000 migrations -- just do rake db:schema:load and have done with it. This is the core team's recommendation, and I think it's a good one.

I have helped other developers who have created several migrations in development, which were applied approximately when created, that were subsequently unable to run when deployed to production. The very same recommendations that guard against this kind of problem will make it possible to run those 1000 migrations (or any subsequence) without problem.

3. I think that db/schema.rb does *NOT* belong in the repository.

Why not? Without it, you can't load the schema with Rake.

This is particularly true in a multi-developer environment where migrations are being initially created on different branches.

I don't see how this makes a difference. Your VCS should be able to merge the schema.rb files. If not, get a better VCS.

Yikes! No! The database itself holds the official version of the schema. If I merge changes from a master branch into my development branch, I will run any new migrations, but I certainly don't want some merge tool to give me a new schema.rb. Depending on the actual content of the migrations on different branches and the order in which they are run, the *actual* schema might be slightly different due to the rules for where new columns are placed on a table.

The database itself isn't going into the repository after all.

That's a red herring.

If you need the file, run a rake db:schema:dump or just run migrations.

No! You need the file so that rake db:schema:load is possible. You've got it completely backwards.

You assume that I need to run db:schema:load, which *I* don't.

If you think about why the migration numbering (file naming) was changed from sequence number to timestamp, you'll realize that the practice of such "interleaved" migrations was a much bigger pain-point than what to do about db/schema.rb.

I don't really understand what you're getting at here.

When migrations were sequentially numbered, two developers on separate branches might both create migration 005 for different purposes. This was a problem. The chance that two developers both create migration 20100422105524 is acceptable small. Of course, they will probably be executed in a different order and if they both add a column to the same table, those columns will likewise be in a different order. If db/schema.rb is in the repository, then lots of commits will have effectively meaningless changes and unless the current HEAD has *my* version is it *not* going to truly represent what's in my database schema.

4. Unless you're initializing (inserting) data via migration, running all the migrations is really not much different than doing a db:schema:load because all the migrations are operating on empty tables.

How can you say this with a straight face? There is no reason *at all* to run lots of migrations rather than doing a simple schema load.

The reason is deploying to an existing production database. You can't do a schema load. The proper way to apply the "new" migrations (which might be kinda old if the last production deploy wasn't so recent) is a db:migrate. As I've stated, you can get into trouble with mismatches between migrations (which don't change after being created and *shouldn't* if created properly) and models (which obviously change over time).

I have projects with 210 and 204 migrations as well as many with fewer. Some of the migrations deal with rather nasty data manipulations to maintain data relationships when the associations are flipped around. It's not something that I would recommend, but the definition of "reasonable" can change dramatically when a client shifts the way he thinks about the data and its evolution.

As an experiment, I set up a new environment for the 204 migration project and ran the migrations from scratch. It takes about 8 minutes. There is about 1 minute of startup time, there are several migrations that load some data including one that takes a bit over 3 minutes to put a few tens of thousands of research datapoints into a set of tables. I'm OK with that amount of time for something as significant as creating a new environment.

(Besides, if you have to "scale up your app", you probably aren't adding a new empty database, but creating a master-slave or sharding for performance.)

Another red herring.

Well, if you can "scale up your app" by starting from an empty schema, go ahead. Perhaps initializing the shards, but then I'd start with the db/schema.rb from production, not something from the repository which almost certainly reflects a development environment however close to production that might be.

Rob, I know you know a lot about the Rails framework, but your advice here will make dealing with databases far more difficult than it needs to be.

If I help someone who recalls some of these nuggets of my wisdom and experience at a time where migrations give them trouble, then I will have made a positive difference.

However, it is that same experience that has led me to the conclusion that keeping db/schema.rb in the source repository is wrong. It is derived data and I would no more put it into the repository than I would have someone put their object files compiled from C or their class files compiled from Java in there.

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

Best, -- Marnen Laibow-Koser http://www.marnen.org marnen@marnen.org --

I'm sure we'll meet again! Take care,

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

12

I think there is a difference between an app like mephisto or similar where many people are going to be deploying instances all over the place and a private app for which there is only ever going to be one deploy. With the latter, the database is the definition of the schema and you don't care about people deploying the app from scratch because no one will

Fred

13

Rob Biedenharn wrote:

Rob Biedenharn wrote: [...]

Matt, Pito, Marnen, and anyone else,

1. The opinion on whether db/schema.rb goes into the source
repository has changed over time.

No. I've used Rails since 1.2.6. Every version has put a comment in the schema.rb file that recommends putting it into version control.

I've been using Rails since 0.13 and I'll restate that this opinion has changed. I can't find it at the moment, but there WAS a version that specifically said in that comment to NOT put the file into source control. At least in 1.2.2, there is no comment either way.

I'll get out my old versions of Rails and look. However, it's pretty much irrelevant, because look what current versions say:

# Note that this schema.rb definition is the authoritative source for your database schema. If you need # to create the application database on another system, you should be using db:schema:load, not running # all the migrations from scratch. The latter is a flawed and unsustainable approach (the more migrations # you'll amass, the slower it'll run and the greater likelihood for issues).

14

Frederick Cheung wrote:

However, it is that same experience that has led me to the conclusion � that keeping db/schema.rb in the source repository is wrong. �It is � derived data and I would no more put it into the repository than I � would have someone put their object files compiled from C or their � class files compiled from Java in there.

I think there is a difference between an app like mephisto or similar where many people are going to be deploying instances all over the place and a private app for which there is only ever going to be one deploy.

I disagree. The severity is different, but the issues are the same.

With the latter, the database is the definition of the schema and you don't care about people deploying the app from scratch because no one will

Two words: staging server.

Fred

Best,

15

"only ever going to be one deploy" ??

When does the app work right the very first time and never has to be changed?

Your second sentence sounds like you've come around to my point: "the database is the definition of the schema" Something goes into the repository in order to track its changes over time, share it with other developers, and hold the definitive copy to be deployed. Things that *don't* go into the repository are things that don't need to be tracked (e.g., log files), aren't or shouldn't be shared (e.g., passwords and private config like database.yml for production), or aren't the definitive copy (e.g., like db/schema.rb unless you're putting it *back* into the repo from the production environment).

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

16

Sorry that wasn't very clear. a better phrasing might be "only one deployment" (which you of course keep working on and updating as needed)

Fred