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/.
|
|
|
|
|
|
|
|
package main
|
|
|
|
|
|
|
|
import (
|
|
|
|
"bytes"
|
2023-08-23 11:39:14 +00:00
|
|
|
"compress/gzip"
|
2018-03-27 11:37:34 +00:00
|
|
|
"context"
|
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
|
|
|
"crypto/tls"
|
2022-01-05 14:17:13 +00:00
|
|
|
"encoding/base64"
|
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
|
|
|
"encoding/json"
|
|
|
|
"encoding/pem"
|
2022-01-05 14:17:13 +00:00
|
|
|
"errors"
|
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
|
|
|
"fmt"
|
2023-08-23 11:39:14 +00:00
|
|
|
io "io"
|
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
|
|
|
"log"
|
|
|
|
"math/rand"
|
|
|
|
"net"
|
|
|
|
"net/http"
|
|
|
|
"net/url"
|
2024-09-10 09:51:07 +00:00
|
|
|
"slices"
|
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
|
|
|
"strconv"
|
2018-01-28 09:24:48 +00:00
|
|
|
"strings"
|
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
|
|
|
"sync"
|
|
|
|
"time"
|
|
|
|
|
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/syncthing/syncthing/internal/gen/discosrv"
|
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/syncthing/syncthing/lib/protocol"
|
2023-09-06 12:36:00 +00:00
|
|
|
"github.com/syncthing/syncthing/lib/stringutil"
|
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
|
|
|
)
|
|
|
|
|
|
|
|
// announcement is the format received from and sent to clients
|
|
|
|
type announcement struct {
|
|
|
|
Seen time.Time `json:"seen"`
|
|
|
|
Addresses []string `json:"addresses"`
|
|
|
|
}
|
|
|
|
|
|
|
|
type apiSrv struct {
|
2024-06-03 05:13:21 +00:00
|
|
|
addr string
|
|
|
|
cert tls.Certificate
|
|
|
|
db database
|
|
|
|
listener net.Listener
|
|
|
|
repl replicator // optional
|
|
|
|
useHTTP bool
|
2024-09-08 06:58:06 +00:00
|
|
|
compression bool
|
|
|
|
gzipWriters sync.Pool
|
2024-09-06 09:14:23 +00:00
|
|
|
seenTracker *retryAfterTracker
|
|
|
|
notSeenTracker *retryAfterTracker
|
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
|
|
|
}
|
|
|
|
|
2024-09-11 09:43:58 +00:00
|
|
|
type replicator interface {
|
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
|
|
|
send(key *protocol.DeviceID, addrs []*discosrv.DatabaseAddress, seen int64)
|
2024-09-11 09:43:58 +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
|
|
|
type requestID int64
|
|
|
|
|
|
|
|
func (i requestID) String() string {
|
|
|
|
return fmt.Sprintf("%016x", int64(i))
|
|
|
|
}
|
|
|
|
|
|
|
|
type contextKey int
|
|
|
|
|
|
|
|
const idKey contextKey = iota
|
|
|
|
|
2024-09-06 09:14:23 +00:00
|
|
|
func newAPISrv(addr string, cert tls.Certificate, db database, repl replicator, useHTTP, compression bool) *apiSrv {
|
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
|
|
|
return &apiSrv{
|
2024-09-06 09:14:23 +00:00
|
|
|
addr: addr,
|
|
|
|
cert: cert,
|
|
|
|
db: db,
|
|
|
|
repl: repl,
|
|
|
|
useHTTP: useHTTP,
|
|
|
|
compression: compression,
|
|
|
|
seenTracker: &retryAfterTracker{
|
|
|
|
name: "seenTracker",
|
|
|
|
bucketStarts: time.Now(),
|
|
|
|
desiredRate: 250,
|
|
|
|
currentDelay: notFoundRetryUnknownMinSeconds,
|
|
|
|
},
|
|
|
|
notSeenTracker: &retryAfterTracker{
|
|
|
|
name: "notSeenTracker",
|
|
|
|
bucketStarts: time.Now(),
|
|
|
|
desiredRate: 250,
|
|
|
|
currentDelay: notFoundRetryUnknownMaxSeconds / 2,
|
|
|
|
},
|
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
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-09-06 09:14:23 +00:00
|
|
|
func (s *apiSrv) Serve(ctx context.Context) error {
|
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 s.useHTTP {
|
|
|
|
listener, err := net.Listen("tcp", s.addr)
|
|
|
|
if err != nil {
|
|
|
|
log.Println("Listen:", err)
|
2020-11-17 12:19:04 +00:00
|
|
|
return 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
|
|
|
}
|
|
|
|
s.listener = listener
|
|
|
|
} else {
|
|
|
|
tlsCfg := &tls.Config{
|
2023-08-23 11:39:52 +00:00
|
|
|
Certificates: []tls.Certificate{s.cert},
|
|
|
|
ClientAuth: tls.RequestClientCert,
|
|
|
|
MinVersion: tls.VersionTLS12,
|
2023-08-30 07:09:50 +00:00
|
|
|
NextProtos: []string{"h2", "http/1.1"},
|
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
|
|
|
}
|
|
|
|
|
|
|
|
tlsListener, err := tls.Listen("tcp", s.addr, tlsCfg)
|
|
|
|
if err != nil {
|
|
|
|
log.Println("Listen:", err)
|
2020-11-17 12:19:04 +00:00
|
|
|
return 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
|
|
|
}
|
|
|
|
s.listener = tlsListener
|
|
|
|
}
|
|
|
|
|
|
|
|
http.HandleFunc("/", s.handler)
|
|
|
|
http.HandleFunc("/ping", handlePing)
|
|
|
|
|
|
|
|
srv := &http.Server{
|
|
|
|
ReadTimeout: httpReadTimeout,
|
|
|
|
WriteTimeout: httpWriteTimeout,
|
|
|
|
MaxHeaderBytes: httpMaxHeaderBytes,
|
2023-08-23 11:39:52 +00:00
|
|
|
ErrorLog: log.New(io.Discard, "", 0),
|
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
|
|
|
}
|
|
|
|
|
2024-09-06 09:14:23 +00:00
|
|
|
go func() {
|
|
|
|
<-ctx.Done()
|
|
|
|
srv.Shutdown(context.Background())
|
|
|
|
}()
|
|
|
|
|
2020-11-17 12:19:04 +00:00
|
|
|
err := srv.Serve(s.listener)
|
|
|
|
if err != nil {
|
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
|
|
|
log.Println("Serve:", err)
|
|
|
|
}
|
2020-11-17 12:19:04 +00:00
|
|
|
return 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
|
|
|
}
|
|
|
|
|
|
|
|
func (s *apiSrv) handler(w http.ResponseWriter, req *http.Request) {
|
|
|
|
t0 := time.Now()
|
|
|
|
|
|
|
|
lw := NewLoggingResponseWriter(w)
|
|
|
|
|
|
|
|
defer func() {
|
|
|
|
diff := time.Since(t0)
|
|
|
|
apiRequestsSeconds.WithLabelValues(req.Method).Observe(diff.Seconds())
|
|
|
|
apiRequestsTotal.WithLabelValues(req.Method, strconv.Itoa(lw.statusCode)).Inc()
|
|
|
|
}()
|
|
|
|
|
|
|
|
reqID := requestID(rand.Int63())
|
2023-08-30 07:36:27 +00:00
|
|
|
req = req.WithContext(context.WithValue(req.Context(), idKey, reqID))
|
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 debug {
|
2023-08-30 07:36:27 +00:00
|
|
|
log.Println(reqID, req.Method, req.URL, req.Proto)
|
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
|
|
|
}
|
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
remoteAddr := &net.TCPAddr{
|
|
|
|
IP: nil,
|
|
|
|
Port: -1,
|
|
|
|
}
|
|
|
|
|
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 s.useHTTP {
|
2023-11-08 11:10:23 +00:00
|
|
|
// X-Forwarded-For can have multiple client IPs; split using the comma separator
|
|
|
|
forwardIP, _, _ := strings.Cut(req.Header.Get("X-Forwarded-For"), ",")
|
|
|
|
|
|
|
|
// net.ParseIP will return nil if leading/trailing whitespace exists; use strings.TrimSpace()
|
|
|
|
remoteAddr.IP = net.ParseIP(strings.TrimSpace(forwardIP))
|
|
|
|
|
2020-06-16 08:58:25 +00:00
|
|
|
if parsedPort, err := strconv.ParseInt(req.Header.Get("X-Client-Port"), 10, 0); err == nil {
|
2020-06-16 07:17:07 +00:00
|
|
|
remoteAddr.Port = int(parsedPort)
|
|
|
|
}
|
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
|
|
|
} else {
|
2020-06-16 07:17:07 +00:00
|
|
|
var err error
|
|
|
|
remoteAddr, err = net.ResolveTCPAddr("tcp", req.RemoteAddr)
|
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 err != nil {
|
|
|
|
log.Println("remoteAddr:", err)
|
|
|
|
lw.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(lw, "Internal Server Error", http.StatusInternalServerError)
|
|
|
|
apiRequestsTotal.WithLabelValues("no_remote_addr").Inc()
|
|
|
|
return
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
switch req.Method {
|
2023-08-30 07:36:27 +00:00
|
|
|
case http.MethodGet:
|
|
|
|
s.handleGET(lw, req)
|
|
|
|
case http.MethodPost:
|
|
|
|
s.handlePOST(remoteAddr, lw, req)
|
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
|
|
|
default:
|
|
|
|
http.Error(lw, "Method Not Allowed", http.StatusMethodNotAllowed)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-08-30 07:36:27 +00:00
|
|
|
func (s *apiSrv) handleGET(w http.ResponseWriter, req *http.Request) {
|
|
|
|
reqID := req.Context().Value(idKey).(requestID)
|
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
|
|
|
|
|
|
|
deviceID, err := protocol.DeviceIDFromString(req.URL.Query().Get("device"))
|
|
|
|
if err != nil {
|
|
|
|
if debug {
|
|
|
|
log.Println(reqID, "bad device param")
|
|
|
|
}
|
|
|
|
lookupRequestsTotal.WithLabelValues("bad_request").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Bad Request", http.StatusBadRequest)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
2024-09-06 09:14:23 +00:00
|
|
|
rec, err := s.db.get(&deviceID)
|
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 err != nil {
|
|
|
|
// some sort of internal error
|
|
|
|
lookupRequestsTotal.WithLabelValues("internal_error").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Internal Server Error", http.StatusInternalServerError)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
if len(rec.Addresses) == 0 {
|
2024-09-06 09:14:23 +00:00
|
|
|
var afterS int
|
|
|
|
if rec.Seen == 0 {
|
|
|
|
afterS = s.notSeenTracker.retryAfterS()
|
|
|
|
lookupRequestsTotal.WithLabelValues("not_found_ever").Inc()
|
|
|
|
} else {
|
|
|
|
afterS = s.seenTracker.retryAfterS()
|
|
|
|
lookupRequestsTotal.WithLabelValues("not_found_recent").Inc()
|
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
|
|
|
}
|
2023-11-03 20:09:19 +00:00
|
|
|
w.Header().Set("Retry-After", strconv.Itoa(afterS))
|
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
|
|
|
http.Error(w, "Not Found", http.StatusNotFound)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
lookupRequestsTotal.WithLabelValues("success").Inc()
|
|
|
|
|
2023-08-23 11:39:14 +00:00
|
|
|
w.Header().Set("Content-Type", "application/json")
|
|
|
|
var bw io.Writer = w
|
|
|
|
|
|
|
|
// Use compression if the client asks for it
|
2024-09-08 06:58:06 +00:00
|
|
|
if s.compression && strings.Contains(req.Header.Get("Accept-Encoding"), "gzip") {
|
|
|
|
gw, ok := s.gzipWriters.Get().(*gzip.Writer)
|
|
|
|
if ok {
|
|
|
|
gw.Reset(w)
|
|
|
|
} else {
|
|
|
|
gw = gzip.NewWriter(w)
|
|
|
|
}
|
2023-08-23 11:39:14 +00:00
|
|
|
w.Header().Set("Content-Encoding", "gzip")
|
|
|
|
defer gw.Close()
|
2024-09-08 06:58:06 +00:00
|
|
|
defer s.gzipWriters.Put(gw)
|
2023-08-23 11:39:14 +00:00
|
|
|
bw = gw
|
|
|
|
}
|
|
|
|
|
|
|
|
json.NewEncoder(bw).Encode(announcement{
|
|
|
|
Seen: time.Unix(0, rec.Seen).Truncate(time.Second),
|
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
|
|
|
Addresses: addressStrs(rec.Addresses),
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2023-08-30 07:36:27 +00:00
|
|
|
func (s *apiSrv) handlePOST(remoteAddr *net.TCPAddr, w http.ResponseWriter, req *http.Request) {
|
|
|
|
reqID := req.Context().Value(idKey).(requestID)
|
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
|
|
|
|
2022-01-05 14:17:13 +00:00
|
|
|
rawCert, err := certificateBytes(req)
|
|
|
|
if err != nil {
|
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 debug {
|
2022-01-05 14:17:13 +00:00
|
|
|
log.Println(reqID, "no certificates:", 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
|
|
|
}
|
|
|
|
announceRequestsTotal.WithLabelValues("no_certificate").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Forbidden", http.StatusForbidden)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
var ann announcement
|
|
|
|
if err := json.NewDecoder(req.Body).Decode(&ann); err != nil {
|
|
|
|
if debug {
|
|
|
|
log.Println(reqID, "decode:", err)
|
|
|
|
}
|
|
|
|
announceRequestsTotal.WithLabelValues("bad_request").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Bad Request", http.StatusBadRequest)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
deviceID := protocol.NewDeviceID(rawCert)
|
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
addresses := fixupAddresses(remoteAddr, ann.Addresses)
|
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 len(addresses) == 0 {
|
|
|
|
announceRequestsTotal.WithLabelValues("bad_request").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Bad Request", http.StatusBadRequest)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
if err := s.handleAnnounce(deviceID, addresses); err != nil {
|
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
|
|
|
announceRequestsTotal.WithLabelValues("internal_error").Inc()
|
|
|
|
w.Header().Set("Retry-After", errorRetryAfterString())
|
|
|
|
http.Error(w, "Internal Server Error", http.StatusInternalServerError)
|
|
|
|
return
|
|
|
|
}
|
|
|
|
|
|
|
|
announceRequestsTotal.WithLabelValues("success").Inc()
|
|
|
|
|
|
|
|
w.Header().Set("Reannounce-After", reannounceAfterString())
|
|
|
|
w.WriteHeader(http.StatusNoContent)
|
|
|
|
}
|
|
|
|
|
|
|
|
func (s *apiSrv) Stop() {
|
|
|
|
s.listener.Close()
|
|
|
|
}
|
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
func (s *apiSrv) handleAnnounce(deviceID protocol.DeviceID, addresses []string) error {
|
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
|
|
|
now := time.Now()
|
|
|
|
expire := now.Add(addressExpiryTime).UnixNano()
|
|
|
|
|
2024-09-15 15:00:32 +00:00
|
|
|
// The address slice must always be sorted for database merges to work
|
|
|
|
// properly.
|
|
|
|
slices.Sort(addresses)
|
|
|
|
addresses = slices.Compact(addresses)
|
|
|
|
|
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
|
|
|
dbAddrs := make([]*discosrv.DatabaseAddress, len(addresses))
|
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
|
|
|
for i := range addresses {
|
|
|
|
dbAddrs[i].Address = addresses[i]
|
|
|
|
dbAddrs[i].Expires = expire
|
|
|
|
}
|
|
|
|
|
|
|
|
seen := now.UnixNano()
|
|
|
|
if s.repl != nil {
|
2024-09-06 09:14:23 +00:00
|
|
|
s.repl.send(&deviceID, dbAddrs, seen)
|
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
|
|
|
}
|
2024-09-06 09:14:23 +00:00
|
|
|
return s.db.merge(&deviceID, dbAddrs, seen)
|
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
|
|
|
}
|
|
|
|
|
2022-07-28 15:17:29 +00:00
|
|
|
func handlePing(w http.ResponseWriter, _ *http.Request) {
|
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
|
|
|
w.WriteHeader(204)
|
|
|
|
}
|
|
|
|
|
2022-01-05 14:17:13 +00:00
|
|
|
func certificateBytes(req *http.Request) ([]byte, error) {
|
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 req.TLS != nil && len(req.TLS.PeerCertificates) > 0 {
|
2022-01-05 14:17:13 +00:00
|
|
|
return req.TLS.PeerCertificates[0].Raw, nil
|
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
|
|
|
}
|
|
|
|
|
2019-10-07 11:55:27 +00:00
|
|
|
var bs []byte
|
|
|
|
|
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 hdr := req.Header.Get("X-SSL-Cert"); hdr != "" {
|
2019-10-07 11:55:27 +00:00
|
|
|
if strings.Contains(hdr, "%") {
|
|
|
|
// Nginx using $ssl_client_escaped_cert
|
|
|
|
// The certificate is in PEM format with url encoding.
|
|
|
|
// We need to decode for the PEM decoder
|
|
|
|
hdr, err := url.QueryUnescape(hdr)
|
|
|
|
if err != nil {
|
|
|
|
// Decoding failed
|
2022-01-05 14:17:13 +00:00
|
|
|
return nil, err
|
2019-10-07 11:55:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
bs = []byte(hdr)
|
|
|
|
} else {
|
|
|
|
// Nginx using $ssl_client_cert
|
|
|
|
// The certificate is in PEM format but with spaces for newlines. We
|
|
|
|
// need to reinstate the newlines for the PEM decoder. But we need to
|
|
|
|
// leave the spaces in the BEGIN and END lines - the first and last
|
|
|
|
// space - alone.
|
|
|
|
bs = []byte(hdr)
|
|
|
|
firstSpace := bytes.Index(bs, []byte(" "))
|
|
|
|
lastSpace := bytes.LastIndex(bs, []byte(" "))
|
|
|
|
for i := firstSpace + 1; i < lastSpace; i++ {
|
|
|
|
if bs[i] == ' ' {
|
|
|
|
bs[i] = '\n'
|
|
|
|
}
|
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
|
|
|
}
|
|
|
|
}
|
2022-01-05 14:17:13 +00:00
|
|
|
} else if hdr := req.Header.Get("X-Tls-Client-Cert-Der-Base64"); hdr != "" {
|
|
|
|
// Caddy {tls_client_certificate_der_base64}
|
|
|
|
hdr, err := base64.StdEncoding.DecodeString(hdr)
|
|
|
|
if err != nil {
|
|
|
|
// Decoding failed
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
|
|
|
|
bs = pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: hdr})
|
2024-09-09 06:24:09 +00:00
|
|
|
} else if cert := req.Header.Get("X-Forwarded-Tls-Client-Cert"); cert != "" {
|
2019-10-07 11:55:27 +00:00
|
|
|
// Traefik 2 passtlsclientcert
|
2023-10-07 02:09:07 +00:00
|
|
|
//
|
|
|
|
// The certificate is in PEM format, maybe with URL encoding
|
|
|
|
// (depends on Traefik version) but without newlines and start/end
|
|
|
|
// statements. We need to decode, reinstate the newlines every 64
|
2019-10-07 11:55:27 +00:00
|
|
|
// character and add statements for the PEM decoder
|
2023-10-07 02:09:07 +00:00
|
|
|
|
2024-09-09 06:24:09 +00:00
|
|
|
if strings.Contains(cert, "%") {
|
|
|
|
if unesc, err := url.QueryUnescape(cert); err == nil {
|
|
|
|
cert = unesc
|
2023-10-14 10:29:53 +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
|
|
|
}
|
2019-10-07 11:55:27 +00:00
|
|
|
|
2024-09-09 06:24:09 +00:00
|
|
|
const (
|
|
|
|
header = "-----BEGIN CERTIFICATE-----"
|
|
|
|
footer = "-----END CERTIFICATE-----"
|
|
|
|
)
|
|
|
|
|
|
|
|
var b bytes.Buffer
|
|
|
|
b.Grow(len(header) + 1 + len(cert) + len(cert)/64 + 1 + len(footer) + 1)
|
|
|
|
|
|
|
|
b.WriteString(header)
|
|
|
|
b.WriteByte('\n')
|
|
|
|
|
|
|
|
for i := 0; i < len(cert); i += 64 {
|
|
|
|
end := i + 64
|
|
|
|
if end > len(cert) {
|
|
|
|
end = len(cert)
|
|
|
|
}
|
|
|
|
b.WriteString(cert[i:end])
|
|
|
|
b.WriteByte('\n')
|
2019-10-07 11:55:27 +00:00
|
|
|
}
|
|
|
|
|
2024-09-09 06:24:09 +00:00
|
|
|
b.WriteString(footer)
|
|
|
|
b.WriteByte('\n')
|
|
|
|
|
|
|
|
bs = b.Bytes()
|
2019-10-07 11:55:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if bs == nil {
|
2022-01-05 14:17:13 +00:00
|
|
|
return nil, errors.New("empty certificate header")
|
2019-10-07 11:55:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
block, _ := pem.Decode(bs)
|
|
|
|
if block == nil {
|
|
|
|
// Decoding failed
|
2022-01-05 14:17:13 +00:00
|
|
|
return nil, errors.New("certificate decode result is empty")
|
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
|
|
|
}
|
|
|
|
|
2022-01-05 14:17:13 +00:00
|
|
|
return block.Bytes, nil
|
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
|
|
|
}
|
|
|
|
|
|
|
|
// fixupAddresses checks the list of addresses, removing invalid ones and
|
|
|
|
// replacing unspecified IPs with the given remote IP.
|
2020-06-16 07:17:07 +00:00
|
|
|
func fixupAddresses(remote *net.TCPAddr, addresses []string) []string {
|
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
|
|
|
fixed := make([]string, 0, len(addresses))
|
|
|
|
for _, annAddr := range addresses {
|
|
|
|
uri, err := url.Parse(annAddr)
|
|
|
|
if err != nil {
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
host, port, err := net.SplitHostPort(uri.Host)
|
|
|
|
if err != nil {
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
ip := net.ParseIP(host)
|
2018-08-30 17:06:35 +00:00
|
|
|
|
|
|
|
// Some classes of IP are no-go.
|
|
|
|
if ip.IsLoopback() || ip.IsMulticast() {
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
2023-11-08 11:10:23 +00:00
|
|
|
if host == "" || ip.IsUnspecified() {
|
|
|
|
if remote != nil {
|
2020-06-16 07:17:07 +00:00
|
|
|
// Replace the unspecified IP with the request source.
|
2018-08-30 17:06:35 +00:00
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
// ... unless the request source is the loopback address or
|
|
|
|
// multicast/unspecified (can't happen, really).
|
2023-11-08 11:10:23 +00:00
|
|
|
if remote.IP == nil || remote.IP.IsLoopback() || remote.IP.IsMulticast() || remote.IP.IsUnspecified() {
|
2020-06-16 07:17:07 +00:00
|
|
|
continue
|
|
|
|
}
|
2018-08-30 17:06:35 +00:00
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
// Do not use IPv6 remote address if requested scheme is ...4
|
|
|
|
// (i.e., tcp4, etc.)
|
|
|
|
if strings.HasSuffix(uri.Scheme, "4") && remote.IP.To4() == nil {
|
|
|
|
continue
|
|
|
|
}
|
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
|
|
|
|
2020-06-16 07:17:07 +00:00
|
|
|
// Do not use IPv4 remote address if requested scheme is ...6
|
|
|
|
if strings.HasSuffix(uri.Scheme, "6") && remote.IP.To4() != nil {
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
host = remote.IP.String()
|
2023-11-08 11:10:23 +00:00
|
|
|
|
|
|
|
} else {
|
|
|
|
// remote is nil, unable to determine host IP
|
|
|
|
continue
|
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
|
|
|
}
|
2023-11-08 11:10:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// If zero port was specified, use remote port.
|
|
|
|
if port == "0" {
|
|
|
|
if remote != nil && remote.Port > 0 {
|
|
|
|
// use remote port
|
2021-11-07 08:58:37 +00:00
|
|
|
port = strconv.Itoa(remote.Port)
|
2023-11-08 11:10:23 +00:00
|
|
|
} else {
|
|
|
|
// unable to determine remote port
|
|
|
|
continue
|
2020-06-16 07:17:07 +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
|
|
|
}
|
|
|
|
|
|
|
|
uri.Host = net.JoinHostPort(host, port)
|
|
|
|
fixed = append(fixed, uri.String())
|
|
|
|
}
|
|
|
|
|
2023-09-06 12:36:00 +00:00
|
|
|
// Remove duplicate addresses
|
|
|
|
fixed = stringutil.UniqueTrimmedStrings(fixed)
|
|
|
|
|
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
|
|
|
return fixed
|
|
|
|
}
|
|
|
|
|
|
|
|
type loggingResponseWriter struct {
|
|
|
|
http.ResponseWriter
|
|
|
|
statusCode int
|
|
|
|
}
|
|
|
|
|
|
|
|
func NewLoggingResponseWriter(w http.ResponseWriter) *loggingResponseWriter {
|
|
|
|
return &loggingResponseWriter{w, http.StatusOK}
|
|
|
|
}
|
|
|
|
|
|
|
|
func (lrw *loggingResponseWriter) WriteHeader(code int) {
|
|
|
|
lrw.statusCode = code
|
|
|
|
lrw.ResponseWriter.WriteHeader(code)
|
|
|
|
}
|
|
|
|
|
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
|
|
|
func addressStrs(dbAddrs []*discosrv.DatabaseAddress) []string {
|
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
|
|
|
res := make([]string, len(dbAddrs))
|
|
|
|
for i, a := range dbAddrs {
|
|
|
|
res[i] = a.Address
|
|
|
|
}
|
|
|
|
return res
|
|
|
|
}
|
|
|
|
|
|
|
|
func errorRetryAfterString() string {
|
|
|
|
return strconv.Itoa(errorRetryAfterSeconds + rand.Intn(errorRetryFuzzSeconds))
|
|
|
|
}
|
|
|
|
|
|
|
|
func reannounceAfterString() string {
|
|
|
|
return strconv.Itoa(reannounceAfterSeconds + rand.Intn(reannounzeFuzzSeconds))
|
|
|
|
}
|
2024-09-06 09:14:23 +00:00
|
|
|
|
|
|
|
type retryAfterTracker struct {
|
|
|
|
name string
|
|
|
|
desiredRate float64 // requests per second
|
|
|
|
|
|
|
|
mut sync.Mutex
|
|
|
|
lastCount int // requests in the last bucket
|
|
|
|
curCount int // requests in the current bucket
|
|
|
|
bucketStarts time.Time // start of the current bucket
|
|
|
|
currentDelay int // current delay in seconds
|
|
|
|
}
|
|
|
|
|
|
|
|
func (t *retryAfterTracker) retryAfterS() int {
|
|
|
|
now := time.Now()
|
|
|
|
t.mut.Lock()
|
|
|
|
if durS := now.Sub(t.bucketStarts).Seconds(); durS > float64(t.currentDelay) {
|
|
|
|
t.bucketStarts = now
|
|
|
|
t.lastCount = t.curCount
|
|
|
|
lastRate := float64(t.lastCount) / durS
|
|
|
|
|
|
|
|
switch {
|
|
|
|
case t.currentDelay > notFoundRetryUnknownMinSeconds &&
|
|
|
|
lastRate < 0.75*t.desiredRate:
|
|
|
|
t.currentDelay = max(8*t.currentDelay/10, notFoundRetryUnknownMinSeconds)
|
|
|
|
case t.currentDelay < notFoundRetryUnknownMaxSeconds &&
|
|
|
|
lastRate > 1.25*t.desiredRate:
|
|
|
|
t.currentDelay = min(3*t.currentDelay/2, notFoundRetryUnknownMaxSeconds)
|
|
|
|
}
|
|
|
|
|
|
|
|
t.curCount = 0
|
|
|
|
}
|
|
|
|
if t.curCount == 0 {
|
|
|
|
retryAfterLevel.WithLabelValues(t.name).Set(float64(t.currentDelay))
|
|
|
|
}
|
|
|
|
t.curCount++
|
|
|
|
t.mut.Unlock()
|
|
|
|
return t.currentDelay + rand.Intn(t.currentDelay/4)
|
|
|
|
}
|