Commit 5a574883 authored by Nick Thomas's avatar Nick Thomas

Remove MySQL references from development docs

I noticed the doc/development/testing_guide/best_practices.md still
referenced the `[run mysql]` tags, etc. They no longer work, so I
removed them, then realised I had better clean up the rest of
doc/development !
parent 95075fee
......@@ -20,7 +20,7 @@ A typical install of GitLab will be on GNU/Linux. It uses Nginx or Apache as a w
We also support deploying GitLab on Kubernetes using our [gitlab Helm chart](https://docs.gitlab.com/charts/).
The GitLab web app uses MySQL or PostgreSQL for persistent database information (e.g. users, permissions, issues, other meta data). GitLab stores the bare git repositories it serves in `/home/git/repositories` by default. It also keeps default branch and hook information with the bare repository.
The GitLab web app uses PostgreSQL for persistent database information (e.g. users, permissions, issues, other meta data). GitLab stores the bare git repositories it serves in `/home/git/repositories` by default. It also keeps default branch and hook information with the bare repository.
When serving repositories over HTTP/HTTPS GitLab utilizes the GitLab API to resolve authorization and access as well as serving git objects.
......@@ -511,7 +511,15 @@ To summarize here's the [directory structure of the `git` user home directory](.
ps aux | grep '^git'
```
GitLab has several components to operate. As a system user (i.e. any user that is not the `git` user) it requires a persistent database (MySQL/PostreSQL) and redis database. It also uses Apache httpd or Nginx to proxypass Unicorn. As the `git` user it starts Sidekiq and Unicorn (a simple ruby HTTP server running on port `8080` by default). Under the GitLab user there are normally 4 processes: `unicorn_rails master` (1 process), `unicorn_rails worker` (2 processes), `sidekiq` (1 process).
GitLab has several components to operate. It requires a persistent database
(PostgreSQL) and redis database, and uses Apache httpd or Nginx to proxypass
Unicorn. All these components should run as different system users to GitLab
(e.g., `postgres`, `redis` and `www-data`, instead of `git`).
As the `git` user it starts Sidekiq and Unicorn (a simple ruby HTTP server
running on port `8080` by default). Under the GitLab user there are normally 4
processes: `unicorn_rails master` (1 process), `unicorn_rails worker`
(2 processes), `sidekiq` (1 process).
### Repository access
......@@ -554,12 +562,9 @@ $ /etc/init.d/nginx
Usage: nginx {start|stop|restart|reload|force-reload|status|configtest}
```
Persistent database (one of the following)
Persistent database
```
/etc/init.d/mysqld
Usage: /etc/init.d/mysqld {start|stop|status|restart|condrestart|try-restart|reload|force-reload}
$ /etc/init.d/postgresql
Usage: /etc/init.d/postgresql {start|stop|restart|reload|force-reload|status} [version ..]
```
......@@ -597,11 +602,6 @@ PostgreSQL
- `/var/log/postgresql/*`
MySQL
- `/var/log/mysql/*`
- `/var/log/mysql.*`
### GitLab specific config files
GitLab has configuration files located in `/home/git/gitlab/config/*`. Commonly referenced config files include:
......
# Hash Indexes
Both PostgreSQL and MySQL support hash indexes besides the regular btree
PostgreSQL supports hash indexes besides the regular btree
indexes. Hash indexes however are to be avoided at all costs. While they may
_sometimes_ provide better performance the cost of rehashing can be very high.
More importantly: at least until PostgreSQL 10.0 hash indexes are not
......
......@@ -9,7 +9,7 @@ bundle exec rake setup
```
The `setup` task is an alias for `gitlab:setup`.
This tasks calls `db:reset` to create the database, calls `add_limits_mysql` that adds limits to the database schema in case of a MySQL database and finally it calls `db:seed_fu` to seed the database.
This tasks calls `db:reset` to create the database, and calls `db:seed_fu` to seed the database.
Note: `db:setup` calls `db:seed` but this does nothing.
### Seeding issues for all or a given project
......
......@@ -2,7 +2,7 @@
Storing SHA1 hashes as strings is not very space efficient. A SHA1 as a string
requires at least 40 bytes, an additional byte to store the encoding, and
perhaps more space depending on the internals of PostgreSQL and MySQL.
perhaps more space depending on the internals of PostgreSQL.
On the other hand, if one were to store a SHA1 as binary one would only need 20
bytes for the actual SHA1, and 1 or 4 bytes of additional space (again depending
......
......@@ -15,14 +15,11 @@ FROM issues
WHERE title LIKE 'WIP:%';
```
On PostgreSQL the `LIKE` statement is case-sensitive. On MySQL this depends on
the case-sensitivity of the collation, which is usually case-insensitive. To
perform a case-insensitive `LIKE` on PostgreSQL you have to use `ILIKE` instead.
This statement in turn isn't supported on MySQL.
On PostgreSQL the `LIKE` statement is case-sensitive. To perform a case-insensitive
`LIKE` you have to use `ILIKE` instead.
To work around this problem you should write `LIKE` queries using Arel instead
of raw SQL fragments as Arel automatically uses `ILIKE` on PostgreSQL and `LIKE`
on MySQL. This means that instead of this:
To handle this automatically you should use `LIKE` queries using Arel instead
of raw SQL fragments, as Arel automatically uses `ILIKE` on PostgreSQL.
```ruby
Issue.where('title LIKE ?', 'WIP:%')
......@@ -45,7 +42,7 @@ table = Issue.arel_table
Issue.where(table[:title].matches('WIP:%').or(table[:foo].matches('WIP:%')))
```
For PostgreSQL this produces:
On PostgreSQL, this produces:
```sql
SELECT *
......@@ -53,18 +50,10 @@ FROM issues
WHERE (title ILIKE 'WIP:%' OR foo ILIKE 'WIP:%')
```
In turn for MySQL this produces:
```sql
SELECT *
FROM issues
WHERE (title LIKE 'WIP:%' OR foo LIKE 'WIP:%')
```
## LIKE & Indexes
Neither PostgreSQL nor MySQL use any indexes when using `LIKE` / `ILIKE` with a
wildcard at the start. For example, this will not use any indexes:
PostgreSQL won't use any indexes when using `LIKE` / `ILIKE` with a wildcard at
the start. For example, this will not use any indexes:
```sql
SELECT *
......@@ -75,9 +64,8 @@ WHERE title ILIKE '%WIP:%';
Because the value for `ILIKE` starts with a wildcard the database is not able to
use an index as it doesn't know where to start scanning the indexes.
MySQL provides no known solution to this problem. Luckily PostgreSQL _does_
provide a solution: trigram GIN indexes. These indexes can be created as
follows:
Luckily, PostgreSQL _does_ provide a solution: trigram GIN indexes. These
indexes can be created as follows:
```sql
CREATE INDEX [CONCURRENTLY] index_name_here
......
......@@ -15,16 +15,6 @@ manifest themselves within our code. When designing our tests, take time to revi
our test design. We can find some helpful heuristics documented in the Handbook in the
[Test Design](https://about.gitlab.com/handbook/engineering/quality/guidelines/test-engineering/test-design/) section.
## Run tests against MySQL
By default, tests are only run against PostgreSQL, but you can run them on
demand against MySQL by following one of the following conventions:
| Convention | Valid example |
|:----------------------|:-----------------------------|
| Include `mysql` in your branch name | `enhance-mysql-support` |
| Include `[run mysql]` in your commit message | `Fix MySQL support<br><br>[run mysql]` |
## Test speed
GitLab has a massive test suite that, without [parallelization], can take hours
......
......@@ -39,7 +39,6 @@ slowest test files and try to improve them.
## CI setup
- On CE and EE, the test suite runs both PostgreSQL and MySQL.
- Rails logging to `log/test.log` is disabled by default in CI [for
performance reasons][logging]. To override this setting, provide the
`RAILS_ENABLE_TEST_LOG` environment variable.
......
......@@ -35,8 +35,8 @@ Once a test is in quarantine, there are 3 choices:
Quarantined tests are run on the CI in dedicated jobs that are allowed to fail:
- `rspec-pg-quarantine` and `rspec-mysql-quarantine` (CE & EE)
- `rspec-pg-quarantine-ee` and `rspec-mysql-quarantine-ee` (EE only)
- `rspec-pg-quarantine` (CE & EE)
- `rspec-pg-quarantine-ee` (EE only)
## Automatic retries and flaky tests detection
......
# Verifying Database Capabilities
Sometimes certain bits of code may only work on a certain database and/or
Sometimes certain bits of code may only work on a certain database
version. While we try to avoid such code as much as possible sometimes it is
necessary to add database (version) specific behaviour.
To facilitate this we have the following methods that you can use:
- `Gitlab::Database.postgresql?`: returns `true` if PostgreSQL is being used
- `Gitlab::Database.mysql?`: returns `true` if MySQL is being used
- `Gitlab::Database.postgresql?`: returns `true` if PostgreSQL is being used.
You can normally just assume this is the case.
- `Gitlab::Database.version`: returns the PostgreSQL version number as a string
in the format `X.Y.Z`. This method does not work for MySQL
in the format `X.Y.Z`.
This allows you to write code such as:
......
......@@ -7,9 +7,8 @@ downtime.
## Adding Columns
On PostgreSQL you can safely add a new column to an existing table as long as it
does **not** have a default value. For example, this query would not require
downtime:
You can safely add a new column to an existing table as long as it does **not**
have a default value. For example, this query would not require downtime:
```sql
ALTER TABLE projects ADD COLUMN random_value int;
......@@ -27,11 +26,6 @@ This requires updating every single row in the `projects` table so that
indexes in a table. This in turn acquires enough locks on the table for it to
effectively block any other queries.
As of MySQL 5.6 adding a column to a table is still quite an expensive
operation, even when using `ALGORITHM=INPLACE` and `LOCK=NONE`. This means
downtime _may_ be required when modifying large tables as otherwise the
operation could potentially take hours to complete.
Adding a column with a default value _can_ be done without requiring downtime
when using the migration helper method
`Gitlab::Database::MigrationHelpers#add_column_with_default`. This method works
......@@ -311,8 +305,7 @@ migrations](background_migrations.md#cleaning-up).
## Adding Indexes
Adding indexes is an expensive process that blocks INSERT and UPDATE queries for
the duration. When using PostgreSQL one can work around this by using the
`CONCURRENTLY` option:
the duration. You can work around this by using the `CONCURRENTLY` option:
```sql
CREATE INDEX CONCURRENTLY index_name ON projects (column_name);
......@@ -336,17 +329,9 @@ end
Note that `add_concurrent_index` can not be reversed automatically, thus you
need to manually define `up` and `down`.
When running this on PostgreSQL the `CONCURRENTLY` option mentioned above is
used. On MySQL this method produces a regular `CREATE INDEX` query.
MySQL doesn't really have a workaround for this. Supposedly it _can_ create
indexes without the need for downtime but only for variable width columns. The
details on this are a bit sketchy. Since it's better to be safe than sorry one
should assume that adding indexes requires downtime on MySQL.
## Dropping Indexes
Dropping an index does not require downtime on both PostgreSQL and MySQL.
Dropping an index does not require downtime.
## Adding Tables
......@@ -370,7 +355,7 @@ transaction this means this approach would require downtime.
GitLab allows you to work around this by using
`Gitlab::Database::MigrationHelpers#add_concurrent_foreign_key`. This method
ensures that when PostgreSQL is used no downtime is needed.
ensures that no downtime is needed.
## Removing Foreign Keys
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment