zitadel/internal/eventstore/local_crdb_test.go

256 lines
6.8 KiB
Go
Raw Normal View History

2020-10-23 14:16:46 +00:00
package eventstore_test
import (
"context"
"database/sql"
"encoding/json"
2020-10-23 14:16:46 +00:00
"os"
"testing"
"time"
2020-10-23 14:16:46 +00:00
"github.com/cockroachdb/cockroach-go/v2/testserver"
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
"github.com/jackc/pgx/v5/pgxpool"
"github.com/jackc/pgx/v5/stdlib"
"github.com/zitadel/logging"
"github.com/zitadel/zitadel/cmd/initialise"
feat(database): support for postgres (#3998) * beginning with postgres statements * try pgx * use pgx * database * init works for postgres * arrays working * init for cockroach * init * start tests * tests * TESTS * ch * ch * chore: use go 1.18 * read stmts * fix typo * tests * connection string * add missing error handler * cleanup * start all apis * go mod tidy * old update * switch back to minute * on conflict * replace string slice with `database.StringArray` in db models * fix tests and start * update go version in dockerfile * setup go * clean up * remove notification migration * update * docs: add deploy guide for postgres * fix: revert sonyflake * use `database.StringArray` for daos * use `database.StringArray` every where * new tables * index naming, metadata primary key, project grant role key type * docs(postgres): change to beta * chore: correct compose * fix(defaults): add empty postgres config * refactor: remove unused code * docs: add postgres to self hosted * fix broken link * so? * change title * add mdx to link * fix stmt * update goreleaser in test-code * docs: improve postgres example * update more projections * fix: add beta log for postgres * revert index name change * prerelease * fix: add sequence to v1 "reduce paniced" * log if nil * add logging * fix: log output * fix(import): check if org exists and user * refactor: imports * fix(user): ignore malformed events * refactor: method naming * fix: test * refactor: correct errors.Is call * ci: don't build dev binaries on main * fix(go releaser): update version to 1.11.0 * fix(user): projection should not break * fix(user): handle error properly * docs: correct config example * Update .releaserc.js * Update .releaserc.js Co-authored-by: Livio Amstutz <livio.a@gmail.com> Co-authored-by: Elio Bischof <eliobischof@gmail.com>
2022-08-31 07:52:43 +00:00
"github.com/zitadel/zitadel/internal/database"
"github.com/zitadel/zitadel/internal/database/cockroach"
"github.com/zitadel/zitadel/internal/eventstore"
es_sql "github.com/zitadel/zitadel/internal/eventstore/repository/sql"
new_es "github.com/zitadel/zitadel/internal/eventstore/v3"
2020-10-23 14:16:46 +00:00
)
var (
testCRDBClient *database.DB
queriers map[string]eventstore.Querier = make(map[string]eventstore.Querier)
pushers map[string]eventstore.Pusher = make(map[string]eventstore.Pusher)
clients map[string]*database.DB = make(map[string]*database.DB)
2020-10-23 14:16:46 +00:00
)
func TestMain(m *testing.M) {
opts := make([]testserver.TestServerOpt, 0, 1)
if version := os.Getenv("ZITADEL_CRDB_VERSION"); version != "" {
opts = append(opts, testserver.CustomVersionOpt(version))
}
ts, err := testserver.NewTestServer(opts...)
2020-10-23 14:16:46 +00:00
if err != nil {
logging.WithFields("error", err).Fatal("unable to start db")
2020-10-23 14:16:46 +00:00
}
testCRDBClient = &database.DB{
Database: new(testDB),
}
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
connConfig, err := pgxpool.ParseConfig(ts.PGURL().String())
if err != nil {
logging.WithFields("error", err).Fatal("unable to parse db url")
}
connConfig.AfterConnect = new_es.RegisterEventstoreTypes
pool, err := pgxpool.NewWithConfig(context.Background(), connConfig)
2020-10-23 14:16:46 +00:00
if err != nil {
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
logging.WithFields("error", err).Fatal("unable to create db pool")
}
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
testCRDBClient.DB = stdlib.OpenDBFromPool(pool)
if err = testCRDBClient.Ping(); err != nil {
logging.WithFields("error", err).Fatal("unable to ping db")
2020-10-23 14:16:46 +00:00
}
v2 := &es_sql.CRDB{DB: testCRDBClient}
queriers["v2(inmemory)"] = v2
clients["v2(inmemory)"] = testCRDBClient
pushers["v3(inmemory)"] = new_es.NewEventstore(testCRDBClient)
clients["v3(inmemory)"] = testCRDBClient
if localDB, err := connectLocalhost(); err == nil {
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
if err = initDB(context.Background(), localDB); err != nil {
logging.WithFields("error", err).Fatal("migrations failed")
}
pushers["v3(singlenode)"] = new_es.NewEventstore(localDB)
clients["v3(singlenode)"] = localDB
}
// pushers["v2(inmemory)"] = v2
2020-10-23 14:16:46 +00:00
defer func() {
testCRDBClient.Close()
ts.Stop()
}()
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
if err = initDB(context.Background(), testCRDBClient); err != nil {
logging.WithFields("error", err).Fatal("migrations failed")
2020-10-23 14:16:46 +00:00
}
os.Exit(m.Run())
}
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
func initDB(ctx context.Context, db *database.DB) error {
feat(database): support for postgres (#3998) * beginning with postgres statements * try pgx * use pgx * database * init works for postgres * arrays working * init for cockroach * init * start tests * tests * TESTS * ch * ch * chore: use go 1.18 * read stmts * fix typo * tests * connection string * add missing error handler * cleanup * start all apis * go mod tidy * old update * switch back to minute * on conflict * replace string slice with `database.StringArray` in db models * fix tests and start * update go version in dockerfile * setup go * clean up * remove notification migration * update * docs: add deploy guide for postgres * fix: revert sonyflake * use `database.StringArray` for daos * use `database.StringArray` every where * new tables * index naming, metadata primary key, project grant role key type * docs(postgres): change to beta * chore: correct compose * fix(defaults): add empty postgres config * refactor: remove unused code * docs: add postgres to self hosted * fix broken link * so? * change title * add mdx to link * fix stmt * update goreleaser in test-code * docs: improve postgres example * update more projections * fix: add beta log for postgres * revert index name change * prerelease * fix: add sequence to v1 "reduce paniced" * log if nil * add logging * fix: log output * fix(import): check if org exists and user * refactor: imports * fix(user): ignore malformed events * refactor: method naming * fix: test * refactor: correct errors.Is call * ci: don't build dev binaries on main * fix(go releaser): update version to 1.11.0 * fix(user): projection should not break * fix(user): handle error properly * docs: correct config example * Update .releaserc.js * Update .releaserc.js Co-authored-by: Livio Amstutz <livio.a@gmail.com> Co-authored-by: Elio Bischof <eliobischof@gmail.com>
2022-08-31 07:52:43 +00:00
initialise.ReadStmts("cockroach")
config := new(database.Config)
config.SetConnector(&cockroach.Config{
User: cockroach.User{
Username: "zitadel",
},
Database: "zitadel",
})
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
err := initialise.Init(ctx, db,
feat(database): support for postgres (#3998) * beginning with postgres statements * try pgx * use pgx * database * init works for postgres * arrays working * init for cockroach * init * start tests * tests * TESTS * ch * ch * chore: use go 1.18 * read stmts * fix typo * tests * connection string * add missing error handler * cleanup * start all apis * go mod tidy * old update * switch back to minute * on conflict * replace string slice with `database.StringArray` in db models * fix tests and start * update go version in dockerfile * setup go * clean up * remove notification migration * update * docs: add deploy guide for postgres * fix: revert sonyflake * use `database.StringArray` for daos * use `database.StringArray` every where * new tables * index naming, metadata primary key, project grant role key type * docs(postgres): change to beta * chore: correct compose * fix(defaults): add empty postgres config * refactor: remove unused code * docs: add postgres to self hosted * fix broken link * so? * change title * add mdx to link * fix stmt * update goreleaser in test-code * docs: improve postgres example * update more projections * fix: add beta log for postgres * revert index name change * prerelease * fix: add sequence to v1 "reduce paniced" * log if nil * add logging * fix: log output * fix(import): check if org exists and user * refactor: imports * fix(user): ignore malformed events * refactor: method naming * fix: test * refactor: correct errors.Is call * ci: don't build dev binaries on main * fix(go releaser): update version to 1.11.0 * fix(user): projection should not break * fix(user): handle error properly * docs: correct config example * Update .releaserc.js * Update .releaserc.js Co-authored-by: Livio Amstutz <livio.a@gmail.com> Co-authored-by: Elio Bischof <eliobischof@gmail.com>
2022-08-31 07:52:43 +00:00
initialise.VerifyUser(config.Username(), ""),
initialise.VerifyDatabase(config.DatabaseName()),
initialise.VerifyGrant(config.DatabaseName(), config.Username()),
initialise.VerifySettings(config.DatabaseName(), config.Username()))
2020-10-23 14:16:46 +00:00
if err != nil {
return err
}
refactor(eventstore): move push logic to sql (#8816) # Which Problems Are Solved If many events are written to the same aggregate id it can happen that zitadel [starts to retry the push transaction](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/eventstore.go#L101) because [the locking behaviour](https://github.com/zitadel/zitadel/blob/48ffc902cc90237d693e7104fc742ee927478da7/internal/eventstore/v3/sequence.go#L25) during push does compute the wrong sequence because newly committed events are not visible to the transaction. These events impact the current sequence. In cases with high command traffic on a single aggregate id this can have severe impact on general performance of zitadel. Because many connections of the `eventstore pusher` database pool are blocked by each other. # How the Problems Are Solved To improve the performance this locking mechanism was removed and the business logic of push is moved to sql functions which reduce network traffic and can be analyzed by the database before the actual push. For clients of the eventstore framework nothing changed. # Additional Changes - after a connection is established prefetches the newly added database types - `eventstore.BaseEvent` now returns the correct revision of the event # Additional Context - part of https://github.com/zitadel/zitadel/issues/8931 --------- Co-authored-by: Tim Möhlmann <tim+github@zitadel.com> Co-authored-by: Livio Spring <livio.a@gmail.com> Co-authored-by: Max Peintner <max@caos.ch> Co-authored-by: Elio Bischof <elio@zitadel.com> Co-authored-by: Stefan Benz <46600784+stebenz@users.noreply.github.com> Co-authored-by: Miguel Cabrerizo <30386061+doncicuto@users.noreply.github.com> Co-authored-by: Joakim Lodén <Loddan@users.noreply.github.com> Co-authored-by: Yxnt <Yxnt@users.noreply.github.com> Co-authored-by: Stefan Benz <stefan@caos.ch> Co-authored-by: Harsha Reddy <harsha.reddy@klaviyo.com> Co-authored-by: Zach H <zhirschtritt@gmail.com>
2024-12-04 13:51:40 +00:00
err = initialise.VerifyZitadel(ctx, db, *config)
if err != nil {
return err
}
// create old events
_, err = db.Exec(oldEventsTable)
return err
}
func connectLocalhost() (*database.DB, error) {
client, err := sql.Open("pgx", "postgresql://root@localhost:26257/defaultdb?sslmode=disable")
if err != nil {
return nil, err
}
if err = client.Ping(); err != nil {
return nil, err
}
return &database.DB{
DB: client,
Database: new(testDB),
}, nil
2020-10-23 14:16:46 +00:00
}
type testDB struct{}
func (_ *testDB) Timetravel(time.Duration) string { return " AS OF SYSTEM TIME '-1 ms' " }
func (*testDB) DatabaseName() string { return "db" }
func (*testDB) Username() string { return "user" }
func (*testDB) Type() string { return "cockroach" }
func generateCommand(aggregateType eventstore.AggregateType, aggregateID string, opts ...func(*testEvent)) eventstore.Command {
e := &testEvent{
BaseEvent: eventstore.BaseEvent{
Agg: &eventstore.Aggregate{
ID: aggregateID,
Type: aggregateType,
ResourceOwner: "ro",
Version: "v1",
},
Service: "svc",
EventType: "test.created",
},
}
for _, opt := range opts {
opt(e)
}
return e
}
type testEvent struct {
eventstore.BaseEvent
uniqueConstraints []*eventstore.UniqueConstraint
}
func (e *testEvent) Payload() any {
return e.BaseEvent.Data
}
func (e *testEvent) UniqueConstraints() []*eventstore.UniqueConstraint {
return e.uniqueConstraints
}
func canceledCtx() context.Context {
ctx, cancel := context.WithCancel(context.Background())
cancel()
return ctx
}
func fillUniqueData(unique_type, field, instanceID string) error {
_, err := testCRDBClient.Exec("INSERT INTO eventstore.unique_constraints (unique_type, unique_field, instance_id) VALUES ($1, $2, $3)", unique_type, field, instanceID)
return err
}
func generateAddUniqueConstraint(table, uniqueField string) func(e *testEvent) {
return func(e *testEvent) {
e.uniqueConstraints = append(e.uniqueConstraints,
&eventstore.UniqueConstraint{
UniqueType: table,
UniqueField: uniqueField,
Action: eventstore.UniqueConstraintAdd,
},
)
}
}
func generateRemoveUniqueConstraint(table, uniqueField string) func(e *testEvent) {
return func(e *testEvent) {
e.uniqueConstraints = append(e.uniqueConstraints,
&eventstore.UniqueConstraint{
UniqueType: table,
UniqueField: uniqueField,
Action: eventstore.UniqueConstraintRemove,
},
)
}
}
func withTestData(data any) func(e *testEvent) {
return func(e *testEvent) {
d, err := json.Marshal(data)
if err != nil {
panic("marshal data failed")
}
e.BaseEvent.Data = d
}
}
func cleanupEventstore(client *database.DB) func() {
return func() {
_, err := client.Exec("TRUNCATE eventstore.events")
if err != nil {
logging.Warnf("unable to truncate events: %v", err)
}
_, err = client.Exec("TRUNCATE eventstore.events2")
if err != nil {
logging.Warnf("unable to truncate events: %v", err)
}
_, err = client.Exec("TRUNCATE eventstore.unique_constraints")
if err != nil {
logging.Warnf("unable to truncate unique constraints: %v", err)
}
}
}
const oldEventsTable = `CREATE TABLE IF NOT EXISTS eventstore.events (
id UUID DEFAULT gen_random_uuid()
, event_type TEXT NOT NULL
, aggregate_type TEXT NOT NULL
, aggregate_id TEXT NOT NULL
, aggregate_version TEXT NOT NULL
, event_sequence BIGINT NOT NULL
, previous_aggregate_sequence BIGINT
, previous_aggregate_type_sequence INT8
, creation_date TIMESTAMPTZ NOT NULL DEFAULT now()
, created_at TIMESTAMPTZ NOT NULL DEFAULT clock_timestamp()
, event_data JSONB
, editor_user TEXT NOT NULL
, editor_service TEXT
, resource_owner TEXT NOT NULL
, instance_id TEXT NOT NULL
, "position" DECIMAL NOT NULL
, in_tx_order INTEGER NOT NULL
, PRIMARY KEY (instance_id, aggregate_type, aggregate_id, event_sequence DESC)
);`