Re: [capistrano] database synchronization on Non-Rails apps
You need a change management system for your database schema. You are thinking in terms of development ease but truly the difficulty lies in preserving production data; there's no way just copying the schema will ever suffice. The moment you need to do an alter table your sunk. You can't be dropping tables just so you can create the schema you need as you can during development. As development marches on you will be faced with situations that need to modify data that won't need to modify schema. You need to manage those changes as well. Rails has the best scheme I've seen and I would guess many frameworks have copied it; what are you writing your app in? A simpler method but requires you to be very deliberate in the queries you write is a pattern that coppermine gallery uses; and probably many other php apps. Consists of two files. An initial SQL file that is designed to create a new schema (seed). The second is for updating an existing one (update). (you may only need update if you aren't creating new databases all the time) When you make a change to you schema you make it to both files being mindful of the context they are run. The layout of the files is the same; a series of properly terminated queries; the trick is this file is read as a string; split on those terminations and run one statement at a time; if it succeeds print "ok"; if it doesn't print "already done"; That's where the deliberate part comes in; since you need to write the queries to be idempotent; ie you should write the query such that you can run it any number of times but it won't damage anything after it's initial run. Since failure is expected; obviously you need to test what you add to the files really well. Check the SQL directory in coppermine to see what I mean. Lastly RUNNING your db schema mgmt system IS in the capistrano perview. But the actual details and implementation of that system is in the perview of your application or application framework. So a forum dedicated to whatever framework you are building your app in would be a better place to get advice for this issue. On Apr 2, 2011, at 1:05 PM, mem wrote: > On 02-04-2011 19:25, Lee Hambley wrote: >> >> Is this correct? >> >> Correct! >> >> (Will reply in depth when I'm home) > > Thanks. > > Well, one question is prevailing. > > I've created a git hook that executes a mysqldump to the schema and saves it > somewhere on a place where the git can reach. > > The git hook seems to be a nice option, because since Capistrano will deploy > committed files, I can rest assured that it will deploy (my last local schema > to a file). > > After reading a bit more, I now understand that this seems to not suffice. > > I mean: "Great I have a sql file with my last schema on the remote server". > Now what? > I still need a way to apply that schema to the remote database. > > If I'm thinking right, could Capistrano help me archiving this last step ? > -- > * You received this message because you are subscribed to the Google Groups > "Capistrano" group. > * To post to this group, send email to capistrano@googlegroups.com > * To unsubscribe from this group, send email to > capistrano+unsubscr...@googlegroups.com For more options, visit this group at > http://groups.google.com/group/capistrano?hl=en -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
Re: [capistrano] database synchronization on Non-Rails apps
On 02-04-2011 19:25, Lee Hambley wrote: Is this correct? Correct! (Will reply in depth when I'm home) Thanks. Well, one question is prevailing. I've created a git hook that executes a mysqldump to the schema and saves it somewhere on a place where the git can reach. The git hook seems to be a nice option, because since Capistrano will deploy committed files, I can rest assured that it will deploy (my last local schema to a file). After reading a bit more, I now understand that this seems to not suffice. I mean: "Great I have a sql file with my last schema on the remote server". Now what? I still need a way to apply that schema to the remote database. If I'm thinking right, could Capistrano help me archiving this last step ? -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
Re: [capistrano] database synchronization on Non-Rails apps
> > Is this correct? > Correct! (Will reply in depth when I'm home) Sent from my iPhone -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
Re: [capistrano] database synchronization on Non-Rails apps
On 02-04-2011 16:55, mem wrote: Can I exclude .rb files to be deployed ? If so, what would the command be? Do I need to exclude .git files or, isn't necessary? Let me guess, since we are using remote_cache and we are using a subversion system, the only thing we have to do is to, take out from our subversion control, the files and folders that we don't want, because, deploy will, on this case, only deploy the committed changes. Is this correct? -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
Re: [capistrano] database synchronization on Non-Rails apps
On 01-04-2011 13:49, Lee Hambley wrote: Márcio, Where can I see some examples about mysql syncronization for a non-rails application? Rails uses a migrations syntax, in a nut-shell, this works like 001-create-table-one.rb, 002-create-table-two.rb, 003-change-table-one.rb, each one is tracked, and if `003` hasn't been run, it gets run… that's all there is to it, they don't preserve data (unless you write them carefully) and it's not a MySQL feature or something. The Rails method works across databes… supporting MySQL, PostgreSQL, etc.. it's a pretty naïve implementation, but it works well. Perhaps your framework gives you something, if not you might want to look at `cap web:disable` and `cap web:enable` - to (also naïeve) tasks for copying a template "maintenance" page into place whilst you do work that requires some downtime (or a few minutes on a quiet DB) Well, I'm not sure if this helps us somehow, but I've put the database schema on a file, using a git hook. Hopping that, by doing this, I can have my local and remote database syncronized (not the data, only schema) and that, I will not need Capistrano there. Not sure of this, since I've nothing tested yet. Regarding assets between deploys, there's a variable called :shared_children - you set it to an array of the `shared between release` directories, this defaults to something like log, config, tmp, etc - these are documented here: https://github.com/capistrano/capistrano/blob/master/lib/capistrano/recipes/deploy.rb#L50 Put whatever directories you need in here, then at first/early deploy time, make sure your assets are in :deploy_to/shared/whatever-directory/ I don't recall writing anything about assets. (my English is so poor that if someone uses synonymous, without pointing them out, I will probably not get (anyway, that's my issue). But, as a bonus, I now understand what shared directory means! 'shared between release' (Ignoring my lack of knowledge in English matters, and if I'm allowed I would called it "preserve"). Anyway, since I use a :remote_cache deploy, I'm deploying only the committed files. This being said, I probably don't need to use the shared directory to deal with "assets". (correct?). If all the above is correct, two questions - taking that I'm using remote_cache method: Can I exclude .rb files to be deployed ? If so, what would the command be? Do I need to exclude .git files or, isn't necessary? Thanks a lot, Márcio -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
Re: [capistrano] database synchronization on Non-Rails apps
Márcio, Where can I see some examples about mysql syncronization for a non-rails > application? Rails uses a migrations syntax, in a nut-shell, this works like 001-create-table-one.rb, 002-create-table-two.rb, 003-change-table-one.rb, each one is tracked, and if `003` hasn't been run, it gets run… that's all there is to it, they don't preserve data (unless you write them carefully) and it's not a MySQL feature or something. The Rails method works across databes… supporting MySQL, PostgreSQL, etc.. it's a pretty naïve implementation, but it works well. Perhaps your framework gives you something, if not you might want to look at `cap web:disable` and `cap web:enable` - to (also naïeve) tasks for copying a template "maintenance" page into place whilst you do work that requires some downtime (or a few minutes on a quiet DB) Regarding assets between deploys, there's a variable called :shared_children - you set it to an array of the `shared between release` directories, this defaults to something like log, config, tmp, etc - these are documented here: https://github.com/capistrano/capistrano/blob/master/lib/capistrano/recipes/deploy.rb#L50 Put whatever directories you need in here, then at first/early deploy time, make sure your assets are in :deploy_to/shared/whatever-directory/ Hope that helps, at least a little (focused questions are more likely to elicit answers out of reluctant mailing list subscribers!). -- * You received this message because you are subscribed to the Google Groups "Capistrano" group. * To post to this group, send email to capistrano@googlegroups.com * To unsubscribe from this group, send email to capistrano+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/capistrano?hl=en
