featurebase/dax/server/test/managed.go
Travis Turner 2843f218bc
Introduce ServiceManager and Refactor DAX Integration tests (#2320)
* Introduce ServiceManager and Refactor DAX Integration tests

The ServiceManager provides an interface with which to manage
featurebase (dax) services (mds, queryer, computer). It replaces the
confusing interface implementations in /dax/server/server.go (which
optionally used pointers to in-process objects to satisfy an interface)
with (for now) http implementations. The thought is that even if we're
running all services in-process, we should communicate between services
over http in order to mirror what we would do in a production
environment where the services are running on different nodes.

This batch of commits does quit a lot, most of which is captured here:

- Added `path` support to `dax.Address`. Address is now a string of the form [scheme]://[host]:[port]/[path].
- Added `Holder.directiveApplied` to determine (in tests) if the computer has completed applying the latest directive. This is somewhat temporary until we improve the mds-to-computer logic.
- Removed the "service prefix" code which was prepending client URL paths with the prefix. Instead, the serviceType (mds, queryer, computer[n] is now part of `dax.Address`).
- Removed, from the dax config, the top level `StorageMethod` and `StorageDSN` and now just have `MDS.Config.DataDir`.
- Added `Computer.Config.N` to specify the number of computers to run in-process.
- Moved the `pilosa.MDS` interface to `computer.Registrar`. This is an example of getting the interfaces defined in the right packages.
- Added `SnapshotTable()` method to the mds client (to align with its API).
- Changed `Balancer.AddJob()` to `Balancer.AddJobs()` to support, for example, adding 256 partitions in a single call. Refactored some of the naive Balancer to account for this.
- Added a `Seed` to the top-level config. It's not really useful because of package `crypto/rand`.
- Added an in-memory implementation of the DisCo interface and disabled etcd in a computer service.
- Create sepearte data-dirs for each in-process computer.
- Disabled grpc in dax.
- Modified the sql3 test definition format to support multiple insert steps and separate query results (to align with those steps).

* Changes necessary to get multiple computer instance running in-process

For now the config looks like this:

```
[computer]
run = true
n = 4
```

but we can probably just change that to be something like:

```
[computer]
run = 4
```

*Issues found running multiple "computers" in-process*
- grpc was trying to bind on the same port
  - changed GRPCListener from `*net.TCPListener` to `net.Listener`
  - created a nopListener and set to that for now (i.e. disabled grpc)
- etcd was starting more than once
  - changed dax to use in-memory implementations of the disco interfaces (i.e. stop using etcd)
- IDAllocator (which uses boltdb) was trying to open the `idalloc.db` file more than once
  - realized we have to set separate data-dirs for each holder. that fixed it.

* Port dax integration tests to ManagedCommand

* Modify Balancer-related methods like AddJob to AddJobs

There were (and still are) a lot of places where we were adding on job
at a time, even when we had a long list of jobs to add. This resulted in
every job add (for example adding 1 of 256 shards) taking ~40ms, or over
10s to create a keyed table. One reason was because each job add was
making multiple boltdb transactions.

* Port over more dax integration test stuff

* Add DirectiveApplied to signify that snapshot/writes have loaded.

We use this in tests to avoid using sleeps.
This should be considered temporary; we're going to need a more robust
solution for determining when a computer node is ready to serve complete
data.

* Finish porting dax integration tests

* Improve godocs

* Remove docker-based DAX integration tests.

* go mod tidy

* Move test/managed.go to avoid package conflicts

* Modify IDK integration tests to work with ServiceManager changes

This is really just computer -> computer0
And the MDS DataDir config change.

* cleanup found during review

* echo $CI_COMMIT_REF_SLUG in CI

* remove docker image arg, use build instead
2022-12-05 14:49:17 -06:00

244 lines
6.6 KiB
Go

// Copyright 2021 Molecula Corp. All rights reserved.
package test
import (
"encoding/json"
"fmt"
"io"
"log"
"net/http"
"os"
"testing"
"time"
"github.com/molecula/featurebase/v3/dax"
computersvc "github.com/molecula/featurebase/v3/dax/computer/service"
"github.com/molecula/featurebase/v3/dax/mds"
mdssvc "github.com/molecula/featurebase/v3/dax/mds/service"
"github.com/molecula/featurebase/v3/dax/queryer"
queryersvc "github.com/molecula/featurebase/v3/dax/queryer/service"
"github.com/molecula/featurebase/v3/dax/server"
"github.com/molecula/featurebase/v3/errors"
fbtest "github.com/molecula/featurebase/v3/test"
"github.com/stretchr/testify/assert"
)
// ManagedCommand represents a test wrapper for server.Command.
type ManagedCommand struct {
*server.Command
svcmgr *dax.ServiceManager
started bool
}
// Manage returns the ServiceManager for the ManagedCommand.
func (mc *ManagedCommand) Manage() *dax.ServiceManager {
return mc.svcmgr
}
// Address returns the advertise address at which the command can be reached.
func (mc *ManagedCommand) Address() dax.Address {
uri := mc.URI()
return dax.Address(uri.String())
}
// Start starts the embedded command.
func (mc *ManagedCommand) Start() error {
if mc.started {
return nil
}
if err := mc.Command.Start(); err != nil {
return errors.Wrap(err, "starting command")
}
mc.started = true
return nil
}
// Close closes the embedded command.
func (mc *ManagedCommand) Close() error {
return mc.Command.Close()
}
// NewMDS adds a new MDSService to the ManagedCommands ServiceManager.
func (mc *ManagedCommand) NewMDS(cfg mds.Config) dax.ServiceKey {
uri := mc.URI()
cfg.Logger = mc.svcmgr.Logger
mc.svcmgr.MDS = mdssvc.New(uri, mds.New(cfg))
return dax.ServicePrefixMDS
}
// NewQueryer adds a new QueryerService to the ManagedCommands ServiceManager.
func (mc *ManagedCommand) NewQueryer(cfg queryer.Config) dax.ServiceKey {
uri := mc.URI()
logger := mc.svcmgr.Logger
cfg.Logger = logger
mc.svcmgr.Queryer = queryersvc.New(uri, queryer.New(cfg), logger)
var mdsAddr dax.Address
if cfg.MDSAddress != "" {
mdsAddr = dax.Address(cfg.MDSAddress + "/" + dax.ServicePrefixMDS)
} else if mc.svcmgr.MDS != nil {
mdsAddr = mc.svcmgr.MDS.Address()
}
// Set MDS
if err := mc.svcmgr.Queryer.SetMDS(mdsAddr); err != nil {
logger.Panicf(errors.Wrap(err, "setting mds").Error())
}
return dax.ServicePrefixQueryer
}
// NewComputer adds a new ComputerService to the ManagedCommands ServiceManager.
func (mc *ManagedCommand) NewComputer() dax.ServiceKey {
cfg := computersvc.CommandConfig{
WriteLoggerRun: mc.Config.WriteLogger.Run,
WriteLoggerConfig: mc.Config.WriteLogger.Config,
SnapshotterRun: mc.Config.Snapshotter.Run,
SnapshotterConfig: mc.Config.Snapshotter.Config,
ComputerConfig: mc.Config.Computer.Config,
RootDataDir: mc.Config.Computer.Config.DataDir,
Stderr: os.Stderr,
Logger: mc.svcmgr.Logger,
}
cfg.ComputerConfig.MDSAddress = mc.svcmgr.MDS.Address().String()
// Add new computer service.
return mc.svcmgr.AddComputer(
computersvc.New(mc.Address(), cfg, cfg.Logger))
}
// Healthy returns true if the provided service's /health endpoint returns 200
// OK. This means that the service has been added to the ServiceManager and
// started, and that its http handler has been dynamically added.
func (mc *ManagedCommand) Healthy(key dax.ServiceKey) bool {
if key == "" {
return false
}
addr := mc.Address()
url := fmt.Sprintf("%s/%s/health", addr.WithScheme("http"), key)
log.Println("HEALTH URL:", url)
res, err := http.Get(url)
if err != nil {
return false
} else if res.StatusCode != http.StatusOK {
return false
}
return true
}
// WaitForApplied is a test helper function which retries a computer's
// /directive endpoint a specified number of times, along with a sleep time in
// between tries, until the computer returns applied=true.
func (mc *ManagedCommand) WaitForApplied(t *testing.T, key dax.ServiceKey, n int, sleep time.Duration) {
t.Helper()
addr := mc.Address()
url := fmt.Sprintf("%s/%s/directive", addr.WithScheme("http"), key)
log.Println("WAIT URL:", url)
for i := 0; i < n; i++ {
resp, err := http.Get(url)
assert.NoError(t, err)
assert.Equal(t, http.StatusOK, resp.StatusCode)
body := resp.Body
defer body.Close()
got := struct {
Applied bool `json:"applied"`
}{}
assert.NoError(t, json.NewDecoder(body).Decode(&got))
if got.Applied {
return
}
t.Logf("Wait (%d/%d): url: %s (sleep: %s)\n", i, n, url, sleep.String())
if i < n-1 {
time.Sleep(sleep)
}
}
// Getting to here means the directive endpoint never returned successfully,
// so we need to stop the test.
t.Fatal("WaitForApplied timed out")
}
// NewManagedCommand returns a new instance of Command.
func NewManagedCommand(tb fbtest.DirCleaner, opts ...server.CommandOption) *ManagedCommand {
path := tb.TempDir()
svcmgr := dax.NewServiceManager()
opts = append(opts, server.OptCommandServiceManager(svcmgr))
mc := &ManagedCommand{}
output := io.Discard
if testing.Verbose() {
output = os.Stderr
}
mc.Command = server.NewCommand(output, opts...)
mc.svcmgr = svcmgr
mc.Config.Bind = "http://localhost:0"
mc.Config.MDS.Config.DataDir = path + "/mds"
mc.Config.Computer.Config.DataDir = path
mc.Config.WriteLogger.Config.DataDir = path + "/wl"
mc.Config.Snapshotter.Config.DataDir = path + "/sn"
return mc
}
// DefaultConfig includes a single instance of each service type.
func DefaultConfig() *server.Config {
cfg := server.NewConfig()
cfg.Verbose = true
cfg.MDS.Run = true
cfg.MDS.Config.RegistrationBatchTimeout = 0
cfg.Queryer.Run = true
cfg.Computer.Run = true
cfg.Computer.N = 1
cfg.WriteLogger.Run = true
cfg.Snapshotter.Run = true
return cfg
}
// MustRunManagedCommand starts an in-process set of Services based on the
// provided configuration. If no configuration is provided, it will use the
// DefaultConfig which consists of one instance of each service type.
func MustRunManagedCommand(tb testing.TB, opts ...server.CommandOption) *ManagedCommand {
// If no opts are passed, use the default configuration which includes a
// single instance of each service type. This is really just meant to keep
// test code a bit cleaner when it's not necessary to have a custom service
// configuration.
var basic bool
if len(opts) == 0 {
opts = []server.CommandOption{
server.OptCommandConfig(DefaultConfig()),
}
basic = true
}
mc := NewManagedCommand(tb, opts...)
if err := mc.Start(); err != nil {
tb.Fatalf("starting managed command: %v", err)
}
if basic {
assert.True(tb, mc.Healthy(dax.ServicePrefixMDS))
assert.True(tb, mc.Healthy(dax.ServicePrefixQueryer))
assert.True(tb, mc.Healthy(dax.ServicePrefixComputer+"0"))
}
return mc
}