Related
I'm in the research phase trying to adopt 2012 Database Projects on an existing small project. I'm a C# developer, not a DBA, so I'm not particularly fluent with best practices. I've been searching google and stackoverflow for a few hours now but I still don't know how to handle some key deployment scenarios properly.
1) Over the course of several development cycles, how do I manage multiple versions of my database? If I have a client on v3 of my database and I want to upgrade them to v8, how do I manage this? We currently manage hand-crafted schema and data migration scripts for every version of our product. Do we still need to do this separately or is there something in the new paradigm that supports or replaces this?
2) If the schema changes in such a way that requires data to be moved around, what is the best way to handle this? I assume some work goes in the Pre-Deployment script to preserve the data and then the Post-Deploy script puts it back in the right place. Is that the way of it or is there something better?
3) Any other advice or guidance on how best to work with these new technologies is also greately appreciated!
UPDATE: My understanding of the problem has grown a little since I originally asked this question and while I came up with a workable solution, it wasn't quite the solution I was hoping for. Here's a rewording of my problem:
The problem I'm having is purely data related. If I have a client on version 1 of my application and I want to upgrade them to version 5 of my application, I would have no problems doing so if their database had no data. I'd simply let SSDT intelligently compare schemas and migrate the database in one shot. Unfortunately clients have data so it's not that simple. Schema changes from version 1 of my application to version 2 to version 3 (etc) all impact data. My current strategy for managing data requires I maintain a script for each version upgrade (1 to 2, 2 to 3, etc). This prevents me from going straight from version 1 of my application to version 5 because I have no data migration script to go straight there. The prospect creating custom upgrade scripts for every client or managing upgrade scripts to go from every version to every greater version is exponentially unmanageable. What I was hoping was that there was some sort of strategy SSDT enables that makes managing the data side of things easier, maybe even as easy as the schema side of things. My recent experience with SSDT has not given me any hope of such a strategy existing but I would love to find out differently.
I've been working on this myself, and I can tell you it's not easy.
First, to address the reply by JT - you cannot dismiss "versions", even with declarative updating mechanics that SSDT has. SSDT does a "pretty decent" job (provided you know all the switches and gotchas) of moving any source schema to any target schema, and it's true that this doesn't require verioning per se, but it has no idea how to manage "data motion" (at least not that i can see!). So, just like DBProj, you left to your own devices in Pre/Post scripts. Because the data motion scripts depend on a known start and end schema state, you cannot avoid versioning the DB. The "data motion" scripts, therefore, must be applied to a versioned snapshot of the schema, which means you cannot arbitrarily update a DB from v1 to v8 and expect the data motion scripts v2 to v8 to work (presumably, you wouldn't need a v1 data motion script).
Sadly, I can't see any mechanism in SSDT publishing that allows me to handle this scenario in an integrated way. That means you'll have to add your own scafolding.
The first trick is to track versions within the database (and SSDT project). I started using a trick in DBProj, and brought it over to SSDT, and after doing some research, it turns out that others are using this too. You can apply a DB Extended Property to the database itself (call it "BuildVersion" or "AppVersion" or something like that), and apply the version value to it. You can then capture this extended property in the SSDT project itself, and SSDT will add it as a script (you can then check the publish option that includes extended properties). I then use SQLCMD variables to identify the source and target versions being applied in the current pass. Once you identify the delta of versions between the source (project snapshot) and target (target db about to be updated), you can find all the snapshots that need to be applied. Sadly, this is tricky to do from inside the SSDT deployment, and you'll probably have to move it to the build or deployment pipeline (we use TFS automated deployments and have custom actions to do this).
The next hurdle is to keep snapshots of the schema with their associated data motion scripts. In this case, it helps to make the scripts as idempotent as possible (meaning, you can rerun the scripts without any ill side-effects). It helps to split scripts that can safely be rerun from scripts that must be executed one time only. We're doing the same thing with static reference data (dictionary or lookup tables) - in other words, we have a library of MERGE scripts (one per table) that keep the reference data in sync, and these scripts are included in the post-deployment scripts (via the SQLCMD :r command). The important thing to note here is that you must execute them in the correct order in case any of these reference tables have FK references to each other. We include them in the main post-deploy script in order, and it helps that we created a tool that generates these scripts for us - it also resolves dependency order. We run this generation tool at the close of a "version" to capture the current state of the static reference data. All your other data motion scripts are basically going to be special-case and most likely will be single-use only. In that case, you can do one of two things: you can use an IF statement against the db build/app version, or you can wipe out the 1 time scripts after creating each snapshot package.
It helps to remember that SSDT will disable FK check constraints and only re-enable them after the post-deployment scripts run. This gives you a chance to populate new non-null fields, for example (by the way, you have to enable the option to generate temporary "smart" defaults for non-null columns to make this work). However, FK check constraints are only disabled for tables that SSDT is recreating because of a schema change. For other cases, you are responsible for ensuring that data motion scripts run in the proper order to avoid check constraints complaints (or you manually have disable/re-enable them in your scripts).
DACPAC can help you because DACPAC is essentially a snapshot. It will contain several XML files describing the schema (similar to the build output of the project), but frozen in time at the moment you create it. You can then use SQLPACKAGE.EXE or the deploy provider to publish that package snapshot. I haven't quite figured out how to use the DACPAC versioning, because it's more tied to "registered" data apps, so we're stuck with our own versioning scheme, but we do put our own version info into the DACPAC filename.
I wish I had a more conclusive and exhasutive example to provide, but we're still working out the issues here too.
One thing that really sucks about SSDT is that unlike DBProj, it's currently not extensible. Although it does a much better job than DBProj at a lot of different things, you can't override its default behavior unless you can find some method inside of pre/post scripts of getting around a problem. One of the issues we're trying to resolve right now is that the default method of recreating a table for updates (CCDR) really stinks when you have tens of millions of records.
-UPDATE: I haven't seen this post in some time, but apparently it's been active lately, so I thought I'd add a couple of important notes: if you are using VS2012, the June 2013 release of SSDT now has a Data Comparison tool built-in, and also provides extensibility points - that is to say, you can now include Build Contributors and Deployment Plan Modifiers for the project.
I haven't really found any more useful information on the subject but I've spent some time getting to know the tools, tinkering and playing, and I think I've come up with some acceptable answers to my question. These aren't necessarily the best answers. I still don't know if there are other mechanisms or best practices to better support these scenarios, but here's what I've come up with:
The Pre- and Post-Deploy scripts for a given version of the database are only used migrate data from the previous version. At the start of every development cycle, the scripts are cleaned out and as development proceeds they get fleshed out with whatever sql is needed to safely migrate data from the previous version to the new one. The one exception here is static data in the database. This data is known at design time and maintains a permanent presence in the Post-Deploy scripts in the form of T-SQL MERGE statements. This helps make it possible to deploy any version of the database to a new environment with just the latest publish script. At the end of every development cycle, a publish script is generated from the previous version to the new one. This script will include generated sql to migrate the schema and the hand crafted deploy scripts. Yes, I know the Publish tool can be used directly against a database but that's not a good option for our clients. I am also aware of dacpac files but I'm not really sure how to use them. The generated publish script seems to be the best option I know for production upgrades.
So to answer my scenarios:
1) To upgrade a database from v3 to v8, I would have to execute the generated publish script for v4, then for v5, then for v6, etc. This is very similar to how we do it now. It's well understood and Database Projects seem to make creating/maintaining these scripts much easier.
2) When the schema changes from underneath data, the Pre- and Post-Deploy scripts are used to migrate the data to where it needs to go for the new version. Affected data is essentially backed-up in the Pre-Deploy script and put back into place in the Post-Deploy script.
3) I'm still looking for advice on how best to work with these tools in these scenarios and others. If I got anything wrong here, or if there are any other gotchas I should be aware of, please let me know! Thanks!
In my experience of using SSDT the notion of version numbers (i.e. v1, v2...vX etc...) for databases kinda goes away. This is because SSDT offers a development paradigm known as declarative database development which loosely means that you tell SSDT what state you want your schema to be in and then let SSDT take responsibility for getting it into that state by comparing against what you already have. In this paradigm the notion of deploying v4 then v5 etc.... goes away.
Your pre and post deployment scripts, as you correctly state, exist for the purposes of managing data.
Hope that helps.
JT
I just wanted to say that this thread so far has been excellent.
I have been wrestling with the exact same concerns and am attempting to tackle this problem in our organization, on a fairly large legacy application. We've begun the process of moving toward SSDT (on a TFS branch) but are at the point where we really need to understand the deployment process, and managing custom migrations, and reference/lookup data, along the way.
To complicate things further, our application is one code-base but can be customized per 'customer', so we have about 190 databases we are dealing with, for this one project, not just 3 or so as is probably normal. We do deployments all the time and even setup new customers fairly often. We rely heavily on PowerShell now with old-school incremental release scripts (and associated scripts to create a new customer at that version). I plan to contribute once we figure this all out but please share whatever else you've learned. I do believe we will end up maintaining custom release scripts per version, but we'll see. The idea about maintaining each script within the project, and including a From and To SqlCmd variable is very interesting. If we did that, we would probably prune along the way, physically deleting the really old upgrade scripts once everybody was past that version.
BTW - Side note - On the topic of minimizing waste, we also just spent a bunch of time figuring out how to automate the enforcement of proper naming/data type conventions for columns, as well as automatic generation for all primary and foreign keys, based on naming conventions, as well as index and check constraints etc. The hardest part was dealing with the 'deviants' that didn't follow the rules. Maybe I'll share that too one day if anyone is interested, but for now, I need to pursue this deployment, migration, and reference data story heavily. Thanks again. It's like you guys were speaking exactly what was in my head and looking for this morning.
Is there any way to have data migrations on production database not to be with SQL?
We are using MigratorDotNet and when we build a new funcionality for the application that changes the scheme of the database and we need to do some data updates we have to do this complex and troublesome SQL statements so the data is consistent on production.
Was wondering if there was another way to do this, what are the bests practices to do this? Any ideas on other possible solutions?
We cannot use something like NHibernate because then we have to keep fixing old migrations when the scheme changes, and that can be error prone.
The trick is to use your migration tool and fold said data manipulation statements into the migrations. We use an extended version of the same thing typically for a few projects and it can definitely handle that trick.
If you're already using a migration tool like Migrator.NET then I'd say you're most of the way there. Complex schema/data changes are just a fact of life in the RDBMS world.
Try mite. It let's you do anything that you can do with sql and use sql to do it but have the ability to ensure your database is on the desired version and not risk executing a script that has already run (or miss a script), leaving your database in a consistent state.
If your developers adopt this. Deployments are a simple mite update and then you know problems are product related or data related (but not schema related).
https://github.com/soitgoes/mite
Let me know what you think. I developed this and have been using this with my team for years with great success.
We're upgrading a large site, opendemocracy.net, from Drupal 4.7 to Drupal 6. Our existing install has a lot of superfluous tables in the database from once-used modules we won't be using: ideally, we wouldn't keep these, or our old blocks, etc. This is because: (a) it'd be nice to keep the database as small as possible, (b) it'd be nice for new blocks to start at ID 1, etc., (c) as clean an install as possible should minimise any problems (for example, our crufty old install has some strange problems, e.g. not letting us create any new blocks).
All we really want to keep is our nodes (plus comments), taxonomies, files, statistics, users, path aliases, and one custom table. I've tried copying the tables that I think govern these (listed at the end of this question) into a fresh 4.7 install and then upgrading from this, and it appears to work. My questions are: would this work? Is it a good idea? Is there a better way to achieve some of the same goals?
Also, what if we needed to do the same with another install, also adding its content to the same new install? We have a separate install for our forums. The following approach seemed to work, but I'm worried it will cause problems when creating new content reaches the new nids I've created (how does Drupal check what nids are free; is there a database variable I should change?):
increment the nids (and vids) in the node and term_nodes tables of the forums install by the highest nid (or vid) in the main install
create a new vocabulary in the main install with the same terms as the Forums vocab in the forums install (at maininstall.com/admin/content/forum - I'm not sure how else to specify this is the forum vocabulary)
change the tids in the term_nodes tables of the forums install to the tid values of terms in this new vocabulary in the main install
insert the contents of the modified node and term_nodes tables of the forums install into the node and term_nodes tables of the main install.
PS: Here are the tables I've found which need copying:
audio, audio_attach, audio_image, audio_metadata, comments, files, node, node_access, node_comment_statistics, node_counter, node_revisions, node_type, od_story, term_data, term_hierarchy, term_node, term_relation, term_synonym, url_alias, users, vocabulary, vocabulary_node_types
I don't quite agree with Jeremy on the first part. You have to be very careful to mash around in your Drupal database. Sometimes things may seem to work, but then later, you find that you have created a lot of problems for yourself, because you delete something in one table that's referenced in another. Also id number of your blocks really wont effect Drupals ability to create new blocks etc. If you do want "pretty" ids, make sure you test things thoroughly though. But considering the cost/benefit, this is definitely not worth it.
Your tactic with creating a new 4.7 and then upgrading from there should work just fine. I would, however, suggest that you first copy your database, into a copy of your current distro and then deactivate and uninstall all the modules you wont be using. This will erase all the data that's associated to these modules, and should also help clean up your tables in case that have altered those you are using. This might not do anything, but it could help clean up the data you are using.
When Drupal create new nodes, it simply does a SQL INSERT, so it's the database that handles the ids. So you don't need to wory about that. The migrate module Jeremy suggests, should be able to help you out transferring data into your new Drupal install. Bit if you would rather write a script, what you propose seems just fine.
Note
As many upgrade guides and upgrade handbook say. When doing a major upgrade, you should always upgrade through all of the major Drupal releases (e.g. 4 to 5, 5 to 6...), until you reach your goal. The reason is that the internal structure and thus the database schema changes in every major release. So you need the changes done in the data-structure in version 5, to successfully upgrade to version 6. I didn't mention this in my original post, as took this for common knowledge, but it might be a good idea to mention none the less.
Another thing worth noticing, is that as Henrik O correctly points out, you can change the AUTO_INCREMENT value in your database. I can't remember 4.x upgrades, but I believe that Drupal will take care of this, in part of the 4-5 upgrade, as it implements serial tables in that release instead of managing this itself. Also whether or not to run a query to alter the AUTO_INCREMENT depends on the database backend, as not all backends use this property. PostgreSQL don't manage serials in that way, so it should automatically, start creating nodes with the correct nids, if you use it for Drupal.
I'd say your basic approach is fine, but you are right to be concerned about things silently breaking. Your main concern should indeed be referential integrity, as the id handling changed since version 4.7. Originally, Drupal did not use the auto increment/serial features of the Database engines, but did its own id generation by means of the sequences table. They switched to serials in Drupal 5, but for some reason that I do not remember they still kept the sequences table until dropping it in version 6.
So if I where in your position, I'd add an intermediate step and upgrade/migrate to a Drupal 5 install first, then do another upgrade to Drupal 6. The reasoning is that the upgrading process is a more or less hand tuned collection of operations that got refined using the input of users having troubles doing it. Since the most error reports came in from users doing '1 version' upgrades only, going the same route should minimize the probability of encountering unexpected errors/conditions.
Also, your post on drupal.org reveals that the forum instance you want to 'merge in' is a Drupal 5 install, so you could do the merge while you are in the intermediate Drupal 5 'phase' of your migration. (Alternatively, I'd upgrade that instance to Drupal 6 separately before merging it into the main instance.)
As for the id adjustment in Drupal >= 5, you'd need to adjust the AUTO_INCREMENT start value for each affected table explicitly. For example in MySQL, issuing:
ALTER TABLE node AUTO_INCREMENT = 5432;
would tell the node table to start setting new serials (ids) starting at 5432 from now on
(NOTE: AUTO_INCREMENT is the MySQL way of handling this, if using PostgreSQL, take a look at the documentation for the serial 'pseudo' type and the accompanying sequence generation mechanism)
Obviously, you'd need to test your new instance thoroughly. Put a heavy focus on testing the insertion of new data as well as updating existing data (nodes, terms, anything you migrated), as this would reveal errors with the referential integrity.
Be thorough and you should be fine - Good Luck :)
Edit: You should also check the 'variables' table entries carefully, as some settings there might contain a reference to an id of a 'standard' table entry (e.g. a vocabulary vid, a term tid or something similar - in your case especially forum_nav_vocabulary and forum_containers)
If your approach to creating a cleaner base install seems to work then stick with it.
For the second part of the question you may consider the migrate module. It is designed to copy from non drupal cms to drupal, but should be able to help you add drupal content to another drupal site.
There is a development_structure.sql inside my /db folder of my rails application (rails 2.3.4, ruby 1.8.7) and I am not sure exactly what it does.
Is it needed for some specific environment? (I think I read somewhere that it's used for tests)
Do I need to add it to my git repository?
This post has been used as a reference by a coworker of mine, but the two answers are not exact or informative enough.
development_structure.sql is a low-level dump of the schema, which is necessary when you start to use proprietary database features - either you want to or not, you're going to use them at some point.
Regarding the question of storing it or not, there's some debate. Here is an informative post: http://www.saturnflyer.com/blog/jim/2010/09/14/always-check-in-schema-rb/.
And my take on this follows.
The objective of the development_structure.sql is to sync, for any given commit, the database structure with the code, without having previous knowledge of the schema structure, that is, without having to rely on a pre-existing state of the schema to get the new one.
In a nutshell, by having a schema structure available, whenever you change branch/commit, you load it directly and forget it.
This is mostly valid for dynamic and "crowded" projects, where different branches have differences in the underlying schema structure.
Without having the schema structure stored, you would need to always use an existing reference schema in your database, and migrate it back or forward every time you change branch/commit; several real-world cases can make this process inefficient (e.g. when another branch doesn't have some migrations you currently have, or some migrations can't be rolled back).
Another problem is automated builds, which suffer from the same problems, and even worse, they can't apply manual changes.
The only downside is that it requires a certain habit, which is, to store it every time you run a migration. Easy to say, but also easy to forget.
I don't say you can't live without development_structure.sql - of course you can.
But if you have it, when changing branch/commit you just load-and-forget; if you don't, you [may] have to go through a series of manual steps.
You should not add it to your git repository.
It is a file created automatically by rails when you run migrations with your database.yml configured to connect against a mysql database.
You can view it as an alternative to schema.rb
I believe you can force rails to create it by adding in your environment.rb:
config.active_record.schema_format = :sql
When present this file is used for example by:
rake db:test:clone_structure
Edit
Relevant section in Ruby On Rails Guides.
http://guides.rubyonrails.org/migrations.html#schema-dumping-and-you
They recommend to check it into source control on the wiki.
I personally like to keep it out of it. I like to be able to run all migrations very quickly. It is for me a good sign. If migrations become slow I feel like I am not in total control of my environment anymore. Slowness in migrations generally means I have a lot of data in my development database which I feel wrong.
However, It seems to be a matter of personal taste nowadays.
Follow your instincts on this one.
It's created when you run a rake task to clone your development database to your test database. The development database is outputted to SQL which is then read in to your test DB. You can safely delete it.
In rails 3, you don't even have to write this line,
config.active_record.schema_format = :sql
You can generate this structure.sql file by simply running the above rake command mentioned above
Migrations are undoubtedly better than just firing up phpMyAdmin and changing the schema willy-nilly (as I did during my php days), but after using them for awhile, I think they're fatally flawed.
Version control is a solved problem. The main function of migrations is to keep a history of changes to your database. But storing a different file for each change is a clumsy way to track them. You don't create a new version of post.rb (or a file representing the delta) when you want to add a new virtual attribute -- why should you create a new migration when you want to add a new non-virtual attribute?
Put another way, just as you check post.rb into version control, why not check schema.rb into version control and make the changes to the file directly?
This is functionally the same as keeping a file for each delta, but it's much easier to work with. My mental model is "I want table X to have such and such columns (or really, I want model X to have such and such properties)" -- why should you have to infer from this how to get there from the existing schema; just open up schema.rb and give table X the right columns!
But even the idea that classes wrap tables is an implementation detail! Why can't I just open up post.rb and say:
Class Post
t.string :title
t.text :body
end
If you went with a model like this, you'd have to make a decision about what to do with existing data. But even then, migrations are overkill -- when you migrate data, you're going to lose fidelity when you use a migration's down method.
Anyway, my question is, even if you can't think of a better way, aren't migrations kind of gross?
why not check schema.rb into version control and make the changes to the file directly?
Because the database itself is not in sync with version control.
For instance, you could be using the head of the source tree. But you're connecting to a database that was defined as some past version, not the version you have checked out. The migrations allow you to upgrade or downgrade the database schema from any version and to any version, incrementally.
But to answer your last question, yes, migrations are kind of gross. They implement a redundant revision control system on top of another revision control system. However, neither of these revision control systems is really in sync with the database.
Just to paraphrase what others have said: migrations allow you to protect the data as your schema evolves. The notion of maintaining a single schema.rb file is attractive only until your app goes into production. Thereafter, you'll need a way to migrate your existing users' data as your schema changes.
There are also data-related issues that are important to consider, which migrations solve.
Say an old version of my schema has a feet and inches column. For efficiency purposes, I want to combine that into just an inches column to make sorting and searching easier.
My migration can combine all of the feet and inches data into the inches column (feet * 12 + inches) while it's updating the database (i.e. just before it removes the feet column)
Obviously this being in a migration makes it automatically work when you later apply the changes to your production database.
As it stands, they're annoying and inadequate but quite possibly the best option we have available to us at present. Quite a few smart people have spent quite a lot of time working on the problem and this, so far, is about the best they've been able to come up with. After about 20 years of mostly hand-coding database version updates, I came very rapidly to appreciate migrations as a major improvement when I found ActiveRecord.
As you say, version control is a solved problem. Up to a point I'd agree: it's very solved for text files in particular, less so for other file types and not really very much at all for resources such as databases.
How do migrations look if you view them as version control deltas for databases? They're the sum of the deltas you have to apply to get a schema from one version to another. I'm not aware that even git, for all its super-powerfulness, can take two schema files and generate the necessary DDL to do that.
As far as declaring table content in the model, I believe that's what DataMapper does (no personal experience). I think there may be some DDL inference capabilities there as well.
"even if you can't think of a better way, aren't migrations kind of gross?"
Yes. But they're less gross than anything else we have. Do please let us know when you've completed the non-gross alternative.
I suppose given "even if you can't think of a better way", then yes, in the grand scheme of things, migrations are kind of gross. So are Ruby, Rails, ORMs, SQL, web apps, ...
Migrations have the (not insignificant) advantage that they exist. Gross-but-exists tends to win out over Pleasant-but-nonexistent. I'm sure there probably are pleasant and nonexistent ways to migrate your data, but I'm not sure what that means. :-)
OK, I'm going to take a wild guess here and say that you're probably working all by yourself. In a group development project the power of each individual to take responsibility for just his/her changes to the database required for the code that developer is writing is much much more important.
The alternative is that larger groups of programmers (e.g. 10-15 Java developers where I work) end up relying on a couple of dedicated full time database administrators to do that along with their other maintenance, optimization, etc. duties.