Timothy Messier
79866a287f
This adds the ability to define hooks that will run prior to a schema migration. The hooks should only be used when a migration would fail due to data being in an invalid state, likely due to a bug in the previous schema definition. For example if a column was missing a constraint, and a new migration is attempting to add the constraint, there might already be values that violate the new constraint. In such cases, we would want the user to make a decision to address the data, and need to provide them with details about what data is problematic in case they want to manually address the problem. To help the user, if there is a sane way to address the data, we should provide an option to automatically "repair" the data, allowing the migration to continue. However, this still should be an explicit opt-in from the user, and we need to clearly describe what the repair option would do to the user so they understand the consequences of opting in. To accomplish this each migration can be provided a hook that contains a CheckFunc, RepairFunc, and RepairDescription. These functions must operate at the point-in-time of the migration immediately preceding the current migration, as as such they should not use and domain packages or repositories, instead they should use direct sql with the sql library. The description is used to explain what the RepairFunc will do. Hooks should be defined in `internal/db/schema/migrations/oss/oss.go` and registered with the schema edition. See the example in `internal/db/schema/manager_example_test.go` for more details.
oplog
oplog is a package for writing operation log (oplog) entries for the purpose of replication and verification of the data stored in the Boundary RDBMS.
Usage
// you must init the ticket in its own transaction. You only need
// to init a ticket once in the database. It doesn't need to happen for
// every connection. Once it's persistent in the database, you can simply Get it.
initTx := db.Begin()
ticketer := &GormTicketer{Tx: initTx}
err = ticketer.InitTicket("users")
// if there's no error, then commit the initialized ticket
initTx.Commit()
userCreate := oplog_test.TestUser{
TestUser: oplog_test.TestUser{
Name: loginName,
},
}
tx := db.Begin()
// write the user to the database
tx.Create(&userCreate)
ticketer = &GormTicketer{Tx: db}
// get a ticket for writing users to the oplog
ticket, err := ticketer.GetTicket("users")
// create an entry for the oplog with the entry's metadata (likely the entry's Scope)
newLogEntry := NewEntry(
"test-users",
[]Metadata{
Metadata{
Key: "deployment",
Value: "amex",
},
Metadata{
Key: "project",
Value: "central-info-systems",
},
},
cipherer, // wrapping.Wrapper
ticketer,
)
// write an entry with N messages (variadic parameter) in the order they were sent to the database
_, err = newLogEntry.WriteEntryWith(
context.Background(),
&GormWriter{tx},
ticket,
&Message{Message: &userCreate, TypeName: "user", OpType: OpType_CREATE_OP},
)
// if there's an error writing the oplog then roll EVERYTHING back
if err != nil {
tx.Rollback()
}
// no err means you can commit all the things.
tx.Commit()
TBD/TODO
We need to discuss and decide how Boundary is going to handle the following oplog things:
- SQL migrations: you'll find the package's SQL migrations under: ./db/schema We need to decide how Boundary will manage migrations across the system and we will likely need to reference this package's migrations somehow.
oplog entry
Example Oplog Entry for the Target Aggregate
┌────────────────────────────────────────────────┐
│ FIFO with []byte buffer │
│ │
│┌─Msg 4────┐┌─Msg 3────┐┌─Msg 2────┐ │
││ ││ ││ │ │
││ Tags ││ Host ││ HostSet │ │
││┌────────┐││┌────────┐││┌────────┐│ │
│││ ││││ ││││ ││ │
│││ ││││ ││││ ││ │
│││ Tag ││││ Host ││││HostSet ││ │
│││protobuf││││protobuf││││protobuf││ │
│││ ││││ ││││ ││ │
│││ ││││ ││││ ││ │
││└────────┘││└────────┘││└────────┘│ │
││┌────────┐││┌────────┐││┌────────┐│ │
│││ ││││ ││││ ││ │
│││typeName││││typeName││││typeName││ │
│││ Tag ││││ Host ││││HostSet ││ │
││└────────┘││└────────┘││└────────┘│ │
││┌────────┐││┌────────┐││┌────────┐│ │
│││ ││││ ││││ ││ │
│││ OpType ││││ OpType ││││ OpType ││ │
│││ Create ││││ Create ││││ Create ││ │
││└────────┘││└────────┘││└────────┘│ │
│└──────────┘└──────────┘└──────────┘ │
└────────────────────────────────────────────────┘
oplog tables
oplog tables:
as the diagram shows, we can split the
oplog_entries into multiple tables if
needed for performance.
┌────────────────┐
│┌───────────────┴┐
││┌───────────────┴┐ ┌────────────────┐
│││ oplog_entries │ │ oplog_ticket │
││├────────────────┤ ├────────────────┤
│││id │╲ │id │
│││aggregate_name │──┼───────┼ │aggregate_name │
└┤│data │╱ │version │
└┤ │ │ │
└────────────────┘ └────────────────┘
┼
│
│
┼
╱│╲
┌────────────────┐
│ oplog_metadata │
├────────────────┤
│id │
│entry_id │
│key │
│value │
└────────────────┘
oplog optimistic locking using tickets
Alice's Database
transaction Bob's transaction
│ │ │
│─────────BEGIN──────────────────▶│ │
│ │◀───────────────BEGIN─────────────────┤
│ ┌───────────┴───────────┐ │
│ │ ticket version:1 │ │
│ └───────────┬───────────┘ │
│ Select oplog-ticket for │ │
├──────────────"Target"──────────▶│ Select oplog-ticket for │
│ │◀──────────────"Target"───────────────┤
│ │ │
│ Write to Tables in │ │
├───────Target Aggregate─────────▶│ │
│ │ Write to Tables in │
│ │◀─────────Target Aggregate────────────┤
│ │ │
│ │ │
│ Update Ticket │ │
├────Version = 2 where───────────▶│ │
│ Version = 1 ┌─────────┴─────────┐ │
│ │ ticket version: 2 │ │
│ └─────────┬─────────┘ │
│ │ Update Ticket │
│ │◀───────────Version = 2 where─────────┤
│ │ Version = 1 ┌────┴──────────┐
│ │ │ update failed │
├────────────Commit──────────────▶│ └────┬──────────┘
│ │ │
│ │ │
│ │ │
│ │◀────────────Rollback─────────────────┤
│ │ │
│ │ │
```