9b9d76f92d
This commit introduces the new go-libp2p-gorpc streaming capabilities for Cluster. The main aim is to work towards heavily reducing memory usage when working with very large pinsets. As a side-effect, it takes the chance to revampt all types for all public methods so that pointers to static what should be static objects are not used anymore. This should heavily reduce heap allocations and GC activity. The main change is that state.List now returns a channel from which to read the pins, rather than pins being all loaded into a huge slice. Things reading pins have been all updated to iterate on the channel rather than on the slice. The full pinset is no longer fully loaded onto memory for things that run regularly like StateSync(). Additionally, the /allocations endpoint of the rest API no longer returns an array of pins, but rather streams json-encoded pin objects directly. This change has extended to the restapi client (which puts pins into a channel as they arrive) and to ipfs-cluster-ctl. There are still pending improvements like StatusAll() calls which should also stream responses, and specially BlockPut calls which should stream blocks directly into IPFS on a single call. These are coming up in future commits.
108 lines
2.4 KiB
Go
108 lines
2.4 KiB
Go
package client
|
|
|
|
import (
|
|
"context"
|
|
"fmt"
|
|
"sync"
|
|
"testing"
|
|
|
|
"github.com/ipfs/ipfs-cluster/api"
|
|
ma "github.com/multiformats/go-multiaddr"
|
|
)
|
|
|
|
func TestFailoverConcurrently(t *testing.T) {
|
|
// Create a load balancing client with 5 empty clients and 5 clients with APIs
|
|
// say we want to retry the request for at most 5 times
|
|
cfgs := make([]*Config, 10)
|
|
|
|
// 5 clients with an invalid api address
|
|
for i := 0; i < 5; i++ {
|
|
maddr, _ := ma.NewMultiaddr("/ip4/127.0.0.1/tcp/0")
|
|
cfgs[i] = &Config{
|
|
APIAddr: maddr,
|
|
DisableKeepAlives: true,
|
|
}
|
|
}
|
|
|
|
// 5 clients with APIs
|
|
for i := 5; i < 10; i++ {
|
|
cfgs[i] = &Config{
|
|
APIAddr: apiMAddr(testAPI(t)),
|
|
DisableKeepAlives: true,
|
|
}
|
|
}
|
|
|
|
// Run many requests at the same time
|
|
|
|
// With Failover strategy, it would go through first 5 empty clients
|
|
// and then 6th working client. Thus, all requests should always succeed.
|
|
testRunManyRequestsConcurrently(t, cfgs, &Failover{}, 200, 6, true)
|
|
// First 5 clients are empty. Thus, all requests should fail.
|
|
testRunManyRequestsConcurrently(t, cfgs, &Failover{}, 200, 5, false)
|
|
}
|
|
|
|
type dummyClient struct {
|
|
defaultClient
|
|
i int
|
|
}
|
|
|
|
// ID returns dummy client's serial number.
|
|
func (d *dummyClient) ID(ctx context.Context) (api.ID, error) {
|
|
return api.ID{
|
|
Peername: fmt.Sprintf("%d", d.i),
|
|
}, nil
|
|
}
|
|
|
|
func TestRoundRobin(t *testing.T) {
|
|
var clients []Client
|
|
// number of clients
|
|
n := 5
|
|
// create n dummy clients
|
|
for i := 0; i < n; i++ {
|
|
c := &dummyClient{
|
|
i: i,
|
|
}
|
|
clients = append(clients, c)
|
|
}
|
|
|
|
roundRobin := loadBalancingClient{
|
|
strategy: &RoundRobin{
|
|
clients: clients,
|
|
length: uint32(len(clients)),
|
|
},
|
|
}
|
|
|
|
// clients should be used in the sequence 1, 2,.., 4, 0.
|
|
for i := 0; i < n; i++ {
|
|
id, _ := roundRobin.ID(context.Background())
|
|
if id.Peername != fmt.Sprintf("%d", (i+1)%n) {
|
|
t.Errorf("clients are not being tried in sequence, expected client: %d, but found: %s", i, id.Peername)
|
|
}
|
|
}
|
|
|
|
}
|
|
|
|
func testRunManyRequestsConcurrently(t *testing.T, cfgs []*Config, strategy LBStrategy, requests int, retries int, pass bool) {
|
|
c, err := NewLBClient(strategy, cfgs, retries)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
var wg sync.WaitGroup
|
|
for i := 0; i < requests; i++ {
|
|
wg.Add(1)
|
|
go func() {
|
|
defer wg.Done()
|
|
ctx := context.Background()
|
|
_, err := c.ID(ctx)
|
|
if err != nil && pass {
|
|
t.Error(err)
|
|
}
|
|
if err == nil && !pass {
|
|
t.Error("request should fail with connection refusal")
|
|
}
|
|
}()
|
|
}
|
|
wg.Wait()
|
|
}
|