syncthing/cmd/stdiscosrv/main.go

171 lines
5.8 KiB
Go
Raw Normal View History

cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// Copyright (C) 2018 The Syncthing Authors.
//
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this file,
// You can obtain one at https://mozilla.org/MPL/2.0/.
2014-06-01 20:50:14 +00:00
2013-12-23 02:35:05 +00:00
package main
import (
2020-11-17 12:19:04 +00:00
"context"
2015-09-13 09:44:33 +00:00
"crypto/tls"
2013-12-23 02:35:05 +00:00
"log"
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
"net/http"
refactor: use modern Protobuf encoder (#9817) At a high level, this is what I've done and why: - I'm moving the protobuf generation for the `protocol`, `discovery` and `db` packages to the modern alternatives, and using `buf` to generate because it's nice and simple. - After trying various approaches on how to integrate the new types with the existing code, I opted for splitting off our own data model types from the on-the-wire generated types. This means we can have a `FileInfo` type with nicer ergonomics and lots of methods, while the protobuf generated type stays clean and close to the wire protocol. It does mean copying between the two when required, which certainly adds a small amount of inefficiency. If we want to walk this back in the future and use the raw generated type throughout, that's possible, this however makes the refactor smaller (!) as it doesn't change everything about the type for everyone at the same time. - I have simply removed in cold blood a significant number of old database migrations. These depended on previous generations of generated messages of various kinds and were annoying to support in the new fashion. The oldest supported database version now is the one from Syncthing 1.9.0 from Sep 7, 2020. - I changed config structs to be regular manually defined structs. For the sake of discussion, some things I tried that turned out not to work... ### Embedding / wrapping Embedding the protobuf generated structs in our existing types as a data container and keeping our methods and stuff: ``` package protocol type FileInfo struct { *generated.FileInfo } ``` This generates a lot of problems because the internal shape of the generated struct is quite different (different names, different types, more pointers), because initializing it doesn't work like you'd expect (i.e., you end up with an embedded nil pointer and a panic), and because the types of child types don't get wrapped. That is, even if we also have a similar wrapper around a `Vector`, that's not the type you get when accessing `someFileInfo.Version`, you get the `*generated.Vector` that doesn't have methods, etc. ### Aliasing ``` package protocol type FileInfo = generated.FileInfo ``` Doesn't help because you can't attach methods to it, plus all the above. ### Generating the types into the target package like we do now and attaching methods This fails because of the different shape of the generated type (as in the embedding case above) plus the generated struct already has a bunch of methods that we can't necessarily override properly (like `String()` and a bunch of getters). ### Methods to functions I considered just moving all the methods we attach to functions in a specific package, so that for example ``` package protocol func (f FileInfo) Equal(other FileInfo) bool ``` would become ``` package fileinfos func Equal(a, b *generated.FileInfo) bool ``` and this would mostly work, but becomes quite verbose and cumbersome, and somewhat limits discoverability (you can't see what methods are available on the type in auto completions, etc). In the end I did this in some cases, like in the database layer where a lot of things like `func (fv *FileVersion) IsEmpty() bool` becomes `func fvIsEmpty(fv *generated.FileVersion)` because they were anyway just internal methods. Fixes #8247
2024-12-01 15:50:17 +00:00
_ "net/http/pprof"
2014-02-20 16:40:15 +00:00
"os"
"os/signal"
"runtime"
"time"
2013-12-23 02:35:05 +00:00
"github.com/alecthomas/kong"
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
"github.com/prometheus/client_golang/prometheus/promhttp"
refactor: use modern Protobuf encoder (#9817) At a high level, this is what I've done and why: - I'm moving the protobuf generation for the `protocol`, `discovery` and `db` packages to the modern alternatives, and using `buf` to generate because it's nice and simple. - After trying various approaches on how to integrate the new types with the existing code, I opted for splitting off our own data model types from the on-the-wire generated types. This means we can have a `FileInfo` type with nicer ergonomics and lots of methods, while the protobuf generated type stays clean and close to the wire protocol. It does mean copying between the two when required, which certainly adds a small amount of inefficiency. If we want to walk this back in the future and use the raw generated type throughout, that's possible, this however makes the refactor smaller (!) as it doesn't change everything about the type for everyone at the same time. - I have simply removed in cold blood a significant number of old database migrations. These depended on previous generations of generated messages of various kinds and were annoying to support in the new fashion. The oldest supported database version now is the one from Syncthing 1.9.0 from Sep 7, 2020. - I changed config structs to be regular manually defined structs. For the sake of discussion, some things I tried that turned out not to work... ### Embedding / wrapping Embedding the protobuf generated structs in our existing types as a data container and keeping our methods and stuff: ``` package protocol type FileInfo struct { *generated.FileInfo } ``` This generates a lot of problems because the internal shape of the generated struct is quite different (different names, different types, more pointers), because initializing it doesn't work like you'd expect (i.e., you end up with an embedded nil pointer and a panic), and because the types of child types don't get wrapped. That is, even if we also have a similar wrapper around a `Vector`, that's not the type you get when accessing `someFileInfo.Version`, you get the `*generated.Vector` that doesn't have methods, etc. ### Aliasing ``` package protocol type FileInfo = generated.FileInfo ``` Doesn't help because you can't attach methods to it, plus all the above. ### Generating the types into the target package like we do now and attaching methods This fails because of the different shape of the generated type (as in the embedding case above) plus the generated struct already has a bunch of methods that we can't necessarily override properly (like `String()` and a bunch of getters). ### Methods to functions I considered just moving all the methods we attach to functions in a specific package, so that for example ``` package protocol func (f FileInfo) Equal(other FileInfo) bool ``` would become ``` package fileinfos func Equal(a, b *generated.FileInfo) bool ``` and this would mostly work, but becomes quite verbose and cumbersome, and somewhat limits discoverability (you can't see what methods are available on the type in auto completions, etc). In the end I did this in some cases, like in the database layer where a lot of things like `func (fv *FileVersion) IsEmpty() bool` becomes `func fvIsEmpty(fv *generated.FileVersion)` because they were anyway just internal methods. Fixes #8247
2024-12-01 15:50:17 +00:00
"github.com/thejerf/suture/v4"
_ "github.com/syncthing/syncthing/lib/automaxprocs"
"github.com/syncthing/syncthing/lib/build"
"github.com/syncthing/syncthing/lib/protocol"
2024-06-03 05:13:21 +00:00
"github.com/syncthing/syncthing/lib/rand"
"github.com/syncthing/syncthing/lib/s3"
"github.com/syncthing/syncthing/lib/tlsutil"
2013-12-23 02:35:05 +00:00
)
2015-11-13 09:14:19 +00:00
const (
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
addressExpiryTime = 2 * time.Hour
databaseStatisticsInterval = 5 * time.Minute
// Reannounce-After is set to reannounceAfterSeconds +
// random(reannounzeFuzzSeconds), similar for Retry-After
reannounceAfterSeconds = 3300
reannounzeFuzzSeconds = 300
errorRetryAfterSeconds = 1500
errorRetryFuzzSeconds = 300
// Retry for not found is notFoundRetrySeenSeconds for records we have
// seen an announcement for (but it's not active right now) and
// notFoundRetryUnknownSeconds for records we have never seen (or not
// seen within the last week).
notFoundRetryUnknownMinSeconds = 60
notFoundRetryUnknownMaxSeconds = 3600
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
httpReadTimeout = 5 * time.Second
httpWriteTimeout = 5 * time.Second
httpMaxHeaderBytes = 1 << 10
// Size of the replication outbox channel
replicationOutboxSize = 10000
2015-11-13 09:14:19 +00:00
)
var debug = false
2013-12-23 02:35:05 +00:00
type CLI struct {
Cert string `group:"Listen" help:"Certificate file" default:"./cert.pem" env:"DISCOVERY_CERT_FILE"`
Key string `group:"Listen" help:"Key file" default:"./key.pem" env:"DISCOVERY_KEY_FILE"`
HTTP bool `group:"Listen" help:"Listen on HTTP (behind an HTTPS proxy)" env:"DISCOVERY_HTTP"`
Compression bool `group:"Listen" help:"Enable GZIP compression of responses" env:"DISCOVERY_COMPRESSION"`
Listen string `group:"Listen" help:"Listen address" default:":8443" env:"DISCOVERY_LISTEN"`
MetricsListen string `group:"Listen" help:"Metrics listen address" env:"DISCOVERY_METRICS_LISTEN"`
DBDir string `group:"Database" help:"Database directory" default:"." env:"DISCOVERY_DB_DIR"`
DBFlushInterval time.Duration `group:"Database" help:"Interval between database flushes" default:"5m" env:"DISCOVERY_DB_FLUSH_INTERVAL"`
DBS3Endpoint string `name:"db-s3-endpoint" group:"Database (S3 backup)" hidden:"true" help:"S3 endpoint for database" env:"DISCOVERY_DB_S3_ENDPOINT"`
DBS3Region string `name:"db-s3-region" group:"Database (S3 backup)" hidden:"true" help:"S3 region for database" env:"DISCOVERY_DB_S3_REGION"`
DBS3Bucket string `name:"db-s3-bucket" group:"Database (S3 backup)" hidden:"true" help:"S3 bucket for database" env:"DISCOVERY_DB_S3_BUCKET"`
DBS3AccessKeyID string `name:"db-s3-access-key-id" group:"Database (S3 backup)" hidden:"true" help:"S3 access key ID for database" env:"DISCOVERY_DB_S3_ACCESS_KEY_ID"`
DBS3SecretKey string `name:"db-s3-secret-key" group:"Database (S3 backup)" hidden:"true" help:"S3 secret key for database" env:"DISCOVERY_DB_S3_SECRET_KEY"`
AMQPAddress string `group:"AMQP replication" hidden:"true" help:"Address to AMQP broker" env:"DISCOVERY_AMQP_ADDRESS"`
Debug bool `short:"d" help:"Print debug output" env:"DISCOVERY_DEBUG"`
Version bool `short:"v" help:"Print version and exit"`
}
2015-03-25 07:16:52 +00:00
func main() {
2015-03-25 07:16:52 +00:00
log.SetOutput(os.Stdout)
var cli CLI
kong.Parse(&cli)
debug = cli.Debug
2014-02-20 16:40:15 +00:00
log.Println(build.LongVersionFor("stdiscosrv"))
if cli.Version {
return
}
buildInfo.WithLabelValues(build.Version, runtime.Version(), build.User, build.Date.UTC().Format("2006-01-02T15:04:05Z")).Set(1)
var cert tls.Certificate
if !cli.HTTP {
var err error
cert, err = tls.LoadX509KeyPair(cli.Cert, cli.Key)
if os.IsNotExist(err) {
log.Println("Failed to load keypair. Generating one, this might take a while...")
cert, err = tlsutil.NewCertificate(cli.Cert, cli.Key, "stdiscosrv", 20*365)
if err != nil {
log.Fatalln("Failed to generate X509 key pair:", err)
}
} else if err != nil {
log.Fatalln("Failed to load keypair:", err)
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
}
devID := protocol.NewDeviceID(cert.Certificate[0])
log.Println("Server device ID is", devID)
}
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// Root of the service tree.
main := suture.New("main", suture.Spec{
PassThroughPanics: true,
Timeout: 2 * time.Minute,
})
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// If configured, use S3 for database backups.
var s3c *s3.Session
if cli.DBS3Endpoint != "" {
var err error
s3c, err = s3.NewSession(cli.DBS3Endpoint, cli.DBS3Region, cli.DBS3Bucket, cli.DBS3AccessKeyID, cli.DBS3SecretKey)
if err != nil {
log.Fatalf("Failed to create S3 session: %v", err)
}
}
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// Start the database.
db := newInMemoryStore(cli.DBDir, cli.DBFlushInterval, s3c)
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
main.Add(db)
// If we have an AMQP broker for replication, start that
var repl replicator
if cli.AMQPAddress != "" {
2024-06-03 05:13:21 +00:00
clientID := rand.String(10)
kr := newAMQPReplicator(cli.AMQPAddress, clientID, db)
2024-06-03 05:13:21 +00:00
main.Add(kr)
repl = kr
2024-06-03 05:13:21 +00:00
}
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// Start the main API server.
qs := newAPISrv(cli.Listen, cert, db, repl, cli.HTTP, cli.Compression)
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
main.Add(qs)
// If we have a metrics port configured, start a metrics handler.
if cli.MetricsListen != "" {
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
go func() {
mux := http.NewServeMux()
mux.Handle("/metrics", promhttp.Handler())
log.Fatal(http.ListenAndServe(cli.MetricsListen, mux))
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
}()
}
2014-04-03 20:44:40 +00:00
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
// Cancel on signal
signalChan := make(chan os.Signal, 1)
signal.Notify(signalChan, os.Interrupt)
go func() {
sig := <-signalChan
log.Printf("Received signal %s; shutting down", sig)
cancel()
}()
cmd/stdiscosrv: New discovery server (fixes #4618) This is a new revision of the discovery server. Relevant changes and non-changes: - Protocol towards clients is unchanged. - Recommended large scale design is still to be deployed nehind nginx (I tested, and it's still a lot faster at terminating TLS). - Database backend is leveldb again, only. It scales enough, is easy to setup, and we don't need any backend to take care of. - Server supports replication. This is a simple TCP channel - protect it with a firewall when deploying over the internet. (We deploy this within the same datacenter, and with firewall.) Any incoming client announces are sent over the replication channel(s) to other peer discosrvs. Incoming replication changes are applied to the database as if they came from clients, but without the TLS/certificate overhead. - Metrics are exposed using the prometheus library, when enabled. - The database values and replication protocol is protobuf, because JSON was quite CPU intensive when I tried that and benchmarked it. - The "Retry-After" value for failed lookups gets slowly increased from a default of 120 seconds, by 5 seconds for each failed lookup, independently by each discosrv. This lowers the query load over time for clients that are never seen. The Retry-After maxes out at 3600 after a couple of weeks of this increase. The number of failed lookups is stored in the database, now and then (avoiding making each lookup a database put). All in all this means clients can be pointed towards a cluster using just multiple A / AAAA records to gain both load sharing and redundancy (if one is down, clients will talk to the remaining ones). GitHub-Pull-Request: https://github.com/syncthing/syncthing/pull/4648
2018-01-14 08:52:31 +00:00
// Engage!
main.Serve(ctx)
2014-09-08 09:48:26 +00:00
}