open-consul/agent/structs/config_entry_intentions.go

846 lines
22 KiB
Go
Raw Normal View History

connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
package structs
import (
"fmt"
"sort"
"strings"
"time"
auto-reload configuration when config files change (#12329) * add config watcher to the config package * add logging to watcher * add test and refactor to add WatcherEvent. * add all API calls and fix a bug with recreated files * add tests for watcher * remove the unnecessary use of context * Add debug log and a test for file rename * use inode to detect if the file is recreated/replaced and only listen to create events. * tidy ups (#1535) * tidy ups * Add tests for inode reconcile * fix linux vs windows syscall * fix linux vs windows syscall * fix windows compile error * increase timeout * use ctime ID * remove remove/creation test as it's a use case that fail in linux * fix linux/windows to use Ino/CreationTime * fix the watcher to only overwrite current file id * fix linter error * fix remove/create test * set reconcile loop to 200 Milliseconds * fix watcher to not trigger event on remove, add more tests * on a remove event try to add the file back to the watcher and trigger the handler if success * fix race condition * fix flaky test * fix race conditions * set level to info * fix when file is removed and get an event for it after * fix to trigger handler when we get a remove but re-add fail * fix error message * add tests for directory watch and fixes * detect if a file is a symlink and return an error on Add * rename Watcher to FileWatcher and remove symlink deref * add fsnotify@v1.5.1 * fix go mod * do not reset timer on errors, rename OS specific files * rename New func * events trigger on write and rename * add missing test * fix flaking tests * fix flaky test * check reconcile when removed * delete invalid file * fix test to create files with different mod time. * back date file instead of sleeping * add watching file in agent command. * fix watcher call to use new API * add configuration and stop watcher when server stop * add certs as watched files * move FileWatcher to the agent start instead of the command code * stop watcher before replacing it * save watched files in agent * add add and remove interfaces to the file watcher * fix remove to not return an error * use `Add` and `Remove` to update certs files * fix tests * close events channel on the file watcher even when the context is done * extract `NotAutoReloadableRuntimeConfig` is a separate struct * fix linter errors * add Ca configs and outgoing verify to the not auto reloadable config * add some logs and fix to use background context * add tests to auto-config reload * remove stale test * add tests to changes to config files * add check to see if old cert files still trigger updates * rename `NotAutoReloadableRuntimeConfig` to `StaticRuntimeConfig` * fix to re add both key and cert file. Add test to cover this case. * review suggestion Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * add check to static runtime config changes * fix test * add changelog file * fix review comments * Apply suggestions from code review Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * update flag description Co-authored-by: FFMMM <FFMMM@users.noreply.github.com> * fix compilation error * add static runtime config support * fix test * fix review comments * fix log test * Update .changelog/12329.txt Co-authored-by: Dan Upton <daniel@floppy.co> * transfer tests to runtime_test.go * fix filewatcher Replace to not deadlock. * avoid having lingering locks Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * split ReloadConfig func * fix warning message Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * convert `FileWatcher` into an interface * fix compilation errors * fix tests * extract func for adding and removing files Co-authored-by: Ashwin Venkatesh <ashwin@hashicorp.com> Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> Co-authored-by: FFMMM <FFMMM@users.noreply.github.com> Co-authored-by: Daniel Upton <daniel@floppy.co>
2022-03-31 19:11:49 +00:00
"github.com/hashicorp/consul/lib/stringslice"
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
"github.com/hashicorp/consul/acl"
)
type ServiceIntentionsConfigEntry struct {
Kind string
Name string // formerly DestinationName
Sources []*SourceIntention
Meta map[string]string `json:",omitempty"` // formerly Intention.Meta
acl.EnterpriseMeta `hcl:",squash" mapstructure:",squash"` // formerly DestinationNS
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
RaftIndex
}
var _ UpdatableConfigEntry = (*ServiceIntentionsConfigEntry)(nil)
func (e *ServiceIntentionsConfigEntry) GetKind() string {
return ServiceIntentions
}
func (e *ServiceIntentionsConfigEntry) GetName() string {
if e == nil {
return ""
}
return e.Name
}
func (e *ServiceIntentionsConfigEntry) GetMeta() map[string]string {
if e == nil {
return nil
}
return e.Meta
}
func (e *ServiceIntentionsConfigEntry) Clone() *ServiceIntentionsConfigEntry {
e2 := *e
e2.Meta = cloneStringStringMap(e.Meta)
e2.Sources = make([]*SourceIntention, len(e.Sources))
for i, src := range e.Sources {
e2.Sources[i] = src.Clone()
}
return &e2
}
func (e *ServiceIntentionsConfigEntry) DestinationServiceName() ServiceName {
return NewServiceName(e.Name, &e.EnterpriseMeta)
}
func (e *ServiceIntentionsConfigEntry) UpdateSourceByLegacyID(legacyID string, update *SourceIntention) bool {
for i, src := range e.Sources {
if src.LegacyID == legacyID {
e.Sources[i] = update
return true
}
}
return false
}
func (e *ServiceIntentionsConfigEntry) UpsertSourceByName(sn ServiceName, upsert *SourceIntention) {
for i, src := range e.Sources {
if src.SourceServiceName() == sn {
e.Sources[i] = upsert
return
}
}
e.Sources = append(e.Sources, upsert)
}
func (e *ServiceIntentionsConfigEntry) DeleteSourceByLegacyID(legacyID string) bool {
for i, src := range e.Sources {
if src.LegacyID == legacyID {
// Delete slice element: https://github.com/golang/go/wiki/SliceTricks#delete
// a = append(a[:i], a[i+1:]...)
e.Sources = append(e.Sources[:i], e.Sources[i+1:]...)
if len(e.Sources) == 0 {
e.Sources = nil
}
return true
}
}
return false
}
func (e *ServiceIntentionsConfigEntry) DeleteSourceByName(sn ServiceName) bool {
for i, src := range e.Sources {
if src.SourceServiceName() == sn {
// Delete slice element: https://github.com/golang/go/wiki/SliceTricks#delete
// a = append(a[:i], a[i+1:]...)
e.Sources = append(e.Sources[:i], e.Sources[i+1:]...)
if len(e.Sources) == 0 {
e.Sources = nil
}
return true
}
}
return false
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
func (e *ServiceIntentionsConfigEntry) ToIntention(src *SourceIntention) *Intention {
meta := e.Meta
if src.LegacyID != "" {
meta = src.LegacyMeta
}
ixn := &Intention{
ID: src.LegacyID,
Description: src.Description,
SourcePartition: src.PartitionOrEmpty(),
SourceNS: src.NamespaceOrDefault(),
SourceName: src.Name,
SourceType: src.Type,
Action: src.Action,
Permissions: src.Permissions,
Meta: meta,
Precedence: src.Precedence,
DestinationPartition: e.PartitionOrEmpty(),
DestinationNS: e.NamespaceOrDefault(),
DestinationName: e.Name,
RaftIndex: e.RaftIndex,
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
if src.LegacyCreateTime != nil {
ixn.CreatedAt = *src.LegacyCreateTime
}
if src.LegacyUpdateTime != nil {
ixn.UpdatedAt = *src.LegacyUpdateTime
}
if src.LegacyID != "" {
// Ensure that pre-1.9.0 secondaries can still replicate legacy
// intentions via the APIs. These require the Hash field to be
// populated.
//
//nolint:staticcheck
ixn.SetHash()
}
return ixn
}
func (e *ServiceIntentionsConfigEntry) LegacyIDFieldsAreAllEmpty() bool {
for _, src := range e.Sources {
if src.LegacyID != "" {
return false
}
}
return true
}
func (e *ServiceIntentionsConfigEntry) LegacyIDFieldsAreAllSet() bool {
for _, src := range e.Sources {
if src.LegacyID == "" {
return false
}
}
return true
}
func (e *ServiceIntentionsConfigEntry) ToIntentions() Intentions {
out := make(Intentions, 0, len(e.Sources))
for _, src := range e.Sources {
out = append(out, e.ToIntention(src))
}
return out
}
type SourceIntention struct {
// Name is the name of the source service. This can be a wildcard "*", but
// only the full value can be a wildcard. Partial wildcards are not
// allowed.
//
// The source may also be a non-Consul service, as specified by SourceType.
//
// formerly Intention.SourceName
Name string
// Action is whether this is an allowlist or denylist intention.
//
// formerly Intention.Action
//
// NOTE: this is mutually exclusive with the Permissions field.
Action IntentionAction `json:",omitempty"`
// Permissions is the list of additional L7 attributes that extend the
// intention definition.
//
// Permissions are interpreted in the order represented in the slice. In
// default-deny mode, deny permissions are logically subtracted from all
// following allow permissions. Multiple allow permissions are then ORed
// together.
//
// For example:
// ["deny /v2/admin", "allow /v2/*", "allow GET /healthz"]
//
// Is logically interpreted as:
// allow: [
// "(/v2/*) AND NOT (/v2/admin)",
// "(GET /healthz) AND NOT (/v2/admin)"
// ]
Permissions []*IntentionPermission `json:",omitempty"`
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
// Precedence is the order that the intention will be applied, with
// larger numbers being applied first. This is a read-only field, on
// any intention update it is updated.
//
// Note we will technically decode this over the wire during a write, but
// we always recompute it on save.
//
// formerly Intention.Precedence
Precedence int
// LegacyID is manipulated just by the bridging code
// used as part of backwards compatibility.
//
// formerly Intention.ID
LegacyID string `json:",omitempty" alias:"legacy_id"`
// Type is the type of the value for the source.
//
// formerly Intention.SourceType
Type IntentionSourceType
// Description is a human-friendly description of this intention.
// It is opaque to Consul and is only stored and transferred in API
// requests.
//
// formerly Intention.Description
Description string `json:",omitempty"`
// LegacyMeta is arbitrary metadata associated with the intention. This is
// opaque to Consul but is served in API responses.
//
// formerly Intention.Meta
LegacyMeta map[string]string `json:",omitempty" alias:"legacy_meta"`
// LegacyCreateTime is formerly Intention.CreatedAt
LegacyCreateTime *time.Time `json:",omitempty" alias:"legacy_create_time"`
// LegacyUpdateTime is formerly Intention.UpdatedAt
LegacyUpdateTime *time.Time `json:",omitempty" alias:"legacy_update_time"`
// Things like L7 rules or Sentinel rules could go here later.
// formerly Intention.SourceNS
acl.EnterpriseMeta `hcl:",squash" mapstructure:",squash"`
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
type IntentionPermission struct {
Action IntentionAction // required: allow|deny
HTTP *IntentionHTTPPermission `json:",omitempty"`
// If we have non-http match criteria for other protocols
// in the future (gRPC, redis, etc) they can go here.
// Support for edge-decoded JWTs would likely be configured
// in a new top level section here.
// If we ever add Sentinel support, this is one place we may
// wish to add it.
}
func (p *IntentionPermission) Clone() *IntentionPermission {
p2 := *p
if p.HTTP != nil {
p2.HTTP = p.HTTP.Clone()
}
return &p2
}
type IntentionHTTPPermission struct {
// PathExact, PathPrefix, and PathRegex are mutually exclusive.
PathExact string `json:",omitempty" alias:"path_exact"`
PathPrefix string `json:",omitempty" alias:"path_prefix"`
PathRegex string `json:",omitempty" alias:"path_regex"`
Header []IntentionHTTPHeaderPermission `json:",omitempty"`
Methods []string `json:",omitempty"`
}
func (p *IntentionHTTPPermission) Clone() *IntentionHTTPPermission {
p2 := *p
if len(p.Header) > 0 {
p2.Header = make([]IntentionHTTPHeaderPermission, 0, len(p.Header))
for _, hdr := range p.Header {
p2.Header = append(p2.Header, hdr)
}
}
auto-reload configuration when config files change (#12329) * add config watcher to the config package * add logging to watcher * add test and refactor to add WatcherEvent. * add all API calls and fix a bug with recreated files * add tests for watcher * remove the unnecessary use of context * Add debug log and a test for file rename * use inode to detect if the file is recreated/replaced and only listen to create events. * tidy ups (#1535) * tidy ups * Add tests for inode reconcile * fix linux vs windows syscall * fix linux vs windows syscall * fix windows compile error * increase timeout * use ctime ID * remove remove/creation test as it's a use case that fail in linux * fix linux/windows to use Ino/CreationTime * fix the watcher to only overwrite current file id * fix linter error * fix remove/create test * set reconcile loop to 200 Milliseconds * fix watcher to not trigger event on remove, add more tests * on a remove event try to add the file back to the watcher and trigger the handler if success * fix race condition * fix flaky test * fix race conditions * set level to info * fix when file is removed and get an event for it after * fix to trigger handler when we get a remove but re-add fail * fix error message * add tests for directory watch and fixes * detect if a file is a symlink and return an error on Add * rename Watcher to FileWatcher and remove symlink deref * add fsnotify@v1.5.1 * fix go mod * do not reset timer on errors, rename OS specific files * rename New func * events trigger on write and rename * add missing test * fix flaking tests * fix flaky test * check reconcile when removed * delete invalid file * fix test to create files with different mod time. * back date file instead of sleeping * add watching file in agent command. * fix watcher call to use new API * add configuration and stop watcher when server stop * add certs as watched files * move FileWatcher to the agent start instead of the command code * stop watcher before replacing it * save watched files in agent * add add and remove interfaces to the file watcher * fix remove to not return an error * use `Add` and `Remove` to update certs files * fix tests * close events channel on the file watcher even when the context is done * extract `NotAutoReloadableRuntimeConfig` is a separate struct * fix linter errors * add Ca configs and outgoing verify to the not auto reloadable config * add some logs and fix to use background context * add tests to auto-config reload * remove stale test * add tests to changes to config files * add check to see if old cert files still trigger updates * rename `NotAutoReloadableRuntimeConfig` to `StaticRuntimeConfig` * fix to re add both key and cert file. Add test to cover this case. * review suggestion Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * add check to static runtime config changes * fix test * add changelog file * fix review comments * Apply suggestions from code review Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * update flag description Co-authored-by: FFMMM <FFMMM@users.noreply.github.com> * fix compilation error * add static runtime config support * fix test * fix review comments * fix log test * Update .changelog/12329.txt Co-authored-by: Dan Upton <daniel@floppy.co> * transfer tests to runtime_test.go * fix filewatcher Replace to not deadlock. * avoid having lingering locks Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * split ReloadConfig func * fix warning message Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> * convert `FileWatcher` into an interface * fix compilation errors * fix tests * extract func for adding and removing files Co-authored-by: Ashwin Venkatesh <ashwin@hashicorp.com> Co-authored-by: R.B. Boyer <4903+rboyer@users.noreply.github.com> Co-authored-by: FFMMM <FFMMM@users.noreply.github.com> Co-authored-by: Daniel Upton <daniel@floppy.co>
2022-03-31 19:11:49 +00:00
p2.Methods = stringslice.CloneStringSlice(p.Methods)
return &p2
}
type IntentionHTTPHeaderPermission struct {
Name string
Present bool `json:",omitempty"`
Exact string `json:",omitempty"`
Prefix string `json:",omitempty"`
Suffix string `json:",omitempty"`
Regex string `json:",omitempty"`
Invert bool `json:",omitempty"`
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
func cloneStringStringMap(m map[string]string) map[string]string {
if m == nil {
return nil
}
m2 := make(map[string]string)
for k, v := range m {
m2[k] = v
}
return m2
}
func (x *SourceIntention) SourceServiceName() ServiceName {
return NewServiceName(x.Name, &x.EnterpriseMeta)
}
func (x *SourceIntention) Clone() *SourceIntention {
x2 := *x
x2.LegacyMeta = cloneStringStringMap(x.LegacyMeta)
if len(x.Permissions) > 0 {
x2.Permissions = make([]*IntentionPermission, 0, len(x.Permissions))
for _, perm := range x.Permissions {
x2.Permissions = append(x2.Permissions, perm.Clone())
}
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
return &x2
}
func (e *ServiceIntentionsConfigEntry) UpdateOver(rawPrev ConfigEntry) error {
if rawPrev == nil {
return nil
}
prev, ok := rawPrev.(*ServiceIntentionsConfigEntry)
if !ok {
return fmt.Errorf("previous config entry is not of type %T: %T", e, rawPrev)
}
var (
prevSourceByName = make(map[ServiceName]*SourceIntention)
prevSourceByLegacyID = make(map[string]*SourceIntention)
)
for _, src := range prev.Sources {
prevSourceByName[src.SourceServiceName()] = src
if src.LegacyID != "" {
prevSourceByLegacyID[src.LegacyID] = src
}
}
for i, src := range e.Sources {
if src.LegacyID == "" {
continue
}
// Check that the LegacyID fields are handled correctly during updates.
if prevSrc, ok := prevSourceByName[src.SourceServiceName()]; ok {
if prevSrc.LegacyID == "" {
return fmt.Errorf("Sources[%d].LegacyID: cannot set this field", i)
} else if src.LegacyID != prevSrc.LegacyID {
return fmt.Errorf("Sources[%d].LegacyID: cannot set this field to a different value", i)
}
}
// Now ensure legacy timestamps carry over properly. We always retain the LegacyCreateTime.
if prevSrc, ok := prevSourceByLegacyID[src.LegacyID]; ok {
if prevSrc.LegacyCreateTime != nil {
// NOTE: we don't want to share the memory here
src.LegacyCreateTime = timePointer(*prevSrc.LegacyCreateTime)
}
}
}
return nil
}
func (e *ServiceIntentionsConfigEntry) Normalize() error {
return e.normalize(false)
}
func (e *ServiceIntentionsConfigEntry) LegacyNormalize() error {
return e.normalize(true)
}
func (e *ServiceIntentionsConfigEntry) normalize(legacyWrite bool) error {
if e == nil {
return fmt.Errorf("config entry is nil")
}
// NOTE: this function must be deterministic so that the raft log doesn't
// diverge. This means no ID assignments or time.Now() usage!
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
e.Kind = ServiceIntentions
e.EnterpriseMeta.Normalize()
for _, src := range e.Sources {
// Default source type
if src.Type == "" {
src.Type = IntentionSourceConsul
}
// If the source namespace is omitted it inherits that of the
// destination.
src.EnterpriseMeta.MergeNoWildcard(&e.EnterpriseMeta)
src.EnterpriseMeta.Normalize()
// Compute the precedence only AFTER normalizing namespaces since the
// namespaces are factored into the calculation.
src.Precedence = computeIntentionPrecedence(e, src)
if legacyWrite {
// We always force meta to be non-nil so that it's an empty map. This
// makes it easy for API responses to not nil-check this everywhere.
if src.LegacyMeta == nil {
src.LegacyMeta = make(map[string]string)
}
} else {
// Legacy fields are cleared, except LegacyMeta which we leave
// populated so that we can later fail the write in Validate() and
// give the user a warning about possible data loss.
src.LegacyID = ""
src.LegacyCreateTime = nil
src.LegacyUpdateTime = nil
}
for _, perm := range src.Permissions {
if perm.HTTP == nil {
continue
}
for j := 0; j < len(perm.HTTP.Methods); j++ {
perm.HTTP.Methods[j] = strings.ToUpper(perm.HTTP.Methods[j])
}
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
// The source intentions closer to the head of the list have higher
// precedence. i.e. index 0 has the highest precedence.
sort.SliceStable(e.Sources, func(i, j int) bool {
return e.Sources[i].Precedence > e.Sources[j].Precedence
})
return nil
}
func timePointer(t time.Time) *time.Time {
if t.IsZero() {
return nil
}
return &t
}
// NOTE: this assumes that the namespaces have been fully normalized.
func computeIntentionPrecedence(entry *ServiceIntentionsConfigEntry, src *SourceIntention) int {
// Max maintains the maximum value that the precedence can be depending
// on the number of exact values in the destination.
var max int
switch intentionCountExact(entry.Name, &entry.EnterpriseMeta) {
case 2:
max = 9
case 1:
max = 6
case 0:
max = 3
default:
// This shouldn't be possible, just set it to zero
return 0
}
// Given the maximum, the exact value is determined based on the
// number of source exact values.
countSrc := intentionCountExact(src.Name, &src.EnterpriseMeta)
return max - (2 - countSrc)
}
// intentionCountExact counts the number of exact values (not wildcards) in
// the given namespace and name.
func intentionCountExact(name string, entMeta *acl.EnterpriseMeta) int {
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
ns := entMeta.NamespaceOrDefault()
// If NS is wildcard, pair must be */* since an exact service cannot follow a wildcard NS
// */* is allowed, but */foo is not
if ns == WildcardSpecifier {
return 0
}
// only the namespace must be exact, since the */* case already returned.
if name == WildcardSpecifier {
return 1
}
return 2
}
func (e *ServiceIntentionsConfigEntry) Validate() error {
return e.validate(false)
}
func (e *ServiceIntentionsConfigEntry) LegacyValidate() error {
return e.validate(true)
}
func (e *ServiceIntentionsConfigEntry) HasWildcardDestination() bool {
dstNS := e.EnterpriseMeta.NamespaceOrDefault()
return dstNS == WildcardSpecifier || e.Name == WildcardSpecifier
}
func (e *ServiceIntentionsConfigEntry) HasAnyPermissions() bool {
for _, src := range e.Sources {
if len(src.Permissions) > 0 {
return true
}
}
return false
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
func (e *ServiceIntentionsConfigEntry) validate(legacyWrite bool) error {
if e.Name == "" {
return fmt.Errorf("Name is required")
}
if err := validateIntentionWildcards(e.Name, &e.EnterpriseMeta); err != nil {
return err
}
destIsWild := e.HasWildcardDestination()
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
if legacyWrite {
if len(e.Meta) > 0 {
return fmt.Errorf("Meta must be omitted for legacy intention writes")
}
} else {
if err := validateConfigEntryMeta(e.Meta); err != nil {
return err
}
}
if len(e.Sources) == 0 {
return fmt.Errorf("At least one source is required")
}
seenSources := make(map[ServiceName]struct{})
for i, src := range e.Sources {
if src.Name == "" {
return fmt.Errorf("Sources[%d].Name is required", i)
}
if err := validateIntentionWildcards(src.Name, &src.EnterpriseMeta); err != nil {
return fmt.Errorf("Sources[%d].%v", i, err)
}
if err := validateSourceIntentionEnterpriseMeta(&src.EnterpriseMeta, &e.EnterpriseMeta); err != nil {
return fmt.Errorf("Sources[%d].%v", i, err)
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
// Length of opaque values
if len(src.Description) > metaValueMaxLength {
return fmt.Errorf(
"Sources[%d].Description exceeds maximum length %d", i, metaValueMaxLength)
}
if legacyWrite {
if len(src.LegacyMeta) > metaMaxKeyPairs {
return fmt.Errorf(
"Sources[%d].Meta exceeds maximum element count %d", i, metaMaxKeyPairs)
}
for k, v := range src.LegacyMeta {
if len(k) > metaKeyMaxLength {
return fmt.Errorf(
"Sources[%d].Meta key %q exceeds maximum length %d",
i, k, metaKeyMaxLength,
)
}
if len(v) > metaValueMaxLength {
return fmt.Errorf(
"Sources[%d].Meta value for key %q exceeds maximum length %d",
i, k, metaValueMaxLength,
)
}
}
if src.LegacyCreateTime == nil {
return fmt.Errorf("Sources[%d].LegacyCreateTime must be set", i)
}
if src.LegacyUpdateTime == nil {
return fmt.Errorf("Sources[%d].LegacyUpdateTime must be set", i)
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
} else {
if len(src.LegacyMeta) > 0 {
return fmt.Errorf("Sources[%d].LegacyMeta must be omitted", i)
}
src.LegacyMeta = nil // ensure it's completely unset
if src.LegacyCreateTime != nil {
return fmt.Errorf("Sources[%d].LegacyCreateTime must be omitted", i)
}
if src.LegacyUpdateTime != nil {
return fmt.Errorf("Sources[%d].LegacyUpdateTime must be omitted", i)
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
if legacyWrite {
if src.LegacyID == "" {
return fmt.Errorf("Sources[%d].LegacyID must be set", i)
}
} else {
if src.LegacyID != "" {
return fmt.Errorf("Sources[%d].LegacyID must be omitted", i)
}
}
if legacyWrite || len(src.Permissions) == 0 {
switch src.Action {
case IntentionActionAllow, IntentionActionDeny:
default:
return fmt.Errorf("Sources[%d].Action must be set to 'allow' or 'deny'", i)
}
}
if len(src.Permissions) > 0 && src.Action != "" {
return fmt.Errorf("Sources[%d].Action must be omitted if Permissions are specified", i)
}
if destIsWild && len(src.Permissions) > 0 {
return fmt.Errorf("Sources[%d].Permissions cannot be specified on intentions with wildcarded destinations", i)
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
switch src.Type {
case IntentionSourceConsul:
default:
return fmt.Errorf("Sources[%d].Type must be set to 'consul'", i)
}
for j, perm := range src.Permissions {
switch perm.Action {
case IntentionActionAllow, IntentionActionDeny:
default:
return fmt.Errorf("Sources[%d].Permissions[%d].Action must be set to 'allow' or 'deny'", i, j)
}
errorPrefix := "Sources[%d].Permissions[%d].HTTP"
if perm.HTTP == nil {
return fmt.Errorf(errorPrefix+" is required", i, j)
}
pathParts := 0
if perm.HTTP.PathExact != "" {
pathParts++
if !strings.HasPrefix(perm.HTTP.PathExact, "/") {
return fmt.Errorf(
errorPrefix+".PathExact doesn't start with '/': %q",
i, j, perm.HTTP.PathExact,
)
}
}
if perm.HTTP.PathPrefix != "" {
pathParts++
if !strings.HasPrefix(perm.HTTP.PathPrefix, "/") {
return fmt.Errorf(
errorPrefix+".PathPrefix doesn't start with '/': %q",
i, j, perm.HTTP.PathPrefix,
)
}
}
if perm.HTTP.PathRegex != "" {
pathParts++
}
if pathParts > 1 {
return fmt.Errorf(
errorPrefix+" should only contain at most one of PathExact, PathPrefix, or PathRegex",
i, j,
)
}
permParts := pathParts
for k, hdr := range perm.HTTP.Header {
if hdr.Name == "" {
return fmt.Errorf(errorPrefix+".Header[%d] missing required Name field", i, j, k)
}
hdrParts := 0
if hdr.Present {
hdrParts++
}
if hdr.Exact != "" {
hdrParts++
}
if hdr.Regex != "" {
hdrParts++
}
if hdr.Prefix != "" {
hdrParts++
}
if hdr.Suffix != "" {
hdrParts++
}
if hdrParts != 1 {
return fmt.Errorf(errorPrefix+".Header[%d] should only contain one of Present, Exact, Prefix, Suffix, or Regex", i, j, k)
}
permParts++
}
if len(perm.HTTP.Methods) > 0 {
found := make(map[string]struct{})
for _, m := range perm.HTTP.Methods {
if !isValidHTTPMethod(m) {
return fmt.Errorf(errorPrefix+".Methods contains an invalid method %q", i, j, m)
}
if _, ok := found[m]; ok {
return fmt.Errorf(errorPrefix+".Methods contains %q more than once", i, j, m)
}
found[m] = struct{}{}
}
permParts++
}
if permParts == 0 {
return fmt.Errorf(errorPrefix+" should not be empty", i, j)
}
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
serviceName := src.SourceServiceName()
if _, exists := seenSources[serviceName]; exists {
return fmt.Errorf("Sources[%d] defines %q more than once", i, serviceName.String())
}
seenSources[serviceName] = struct{}{}
}
return nil
}
// Wildcard usage verification
func validateIntentionWildcards(name string, entMeta *acl.EnterpriseMeta) error {
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
ns := entMeta.NamespaceOrDefault()
if ns != WildcardSpecifier {
if strings.Contains(ns, WildcardSpecifier) {
return fmt.Errorf("Namespace: wildcard character '*' cannot be used with partial values")
}
}
if name != WildcardSpecifier {
if strings.Contains(name, WildcardSpecifier) {
return fmt.Errorf("Name: wildcard character '*' cannot be used with partial values")
}
if ns == WildcardSpecifier {
return fmt.Errorf("Name: exact value cannot follow wildcard namespace")
}
}
if strings.Contains(entMeta.PartitionOrDefault(), WildcardSpecifier) {
return fmt.Errorf("Partition: cannot use wildcard '*' in partition")
}
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
return nil
}
func (e *ServiceIntentionsConfigEntry) GetRaftIndex() *RaftIndex {
if e == nil {
return &RaftIndex{}
}
return &e.RaftIndex
}
func (e *ServiceIntentionsConfigEntry) GetEnterpriseMeta() *acl.EnterpriseMeta {
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
if e == nil {
return nil
}
return &e.EnterpriseMeta
}
func (e *ServiceIntentionsConfigEntry) CanRead(authz acl.Authorizer) error {
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
var authzContext acl.AuthorizerContext
e.FillAuthzContext(&authzContext)
return authz.ToAllowAuthorizer().IntentionReadAllowed(e.GetName(), &authzContext)
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
func (e *ServiceIntentionsConfigEntry) CanWrite(authz acl.Authorizer) error {
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
var authzContext acl.AuthorizerContext
e.FillAuthzContext(&authzContext)
return authz.ToAllowAuthorizer().IntentionWriteAllowed(e.GetName(), &authzContext)
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
}
func MigrateIntentions(ixns Intentions) []*ServiceIntentionsConfigEntry {
if len(ixns) == 0 {
return nil
}
collated := make(map[ServiceName]*ServiceIntentionsConfigEntry)
for _, ixn := range ixns {
thisEntry := ixn.ToConfigEntry(true)
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
sn := thisEntry.DestinationServiceName()
if entry, ok := collated[sn]; ok {
entry.Sources = append(entry.Sources, thisEntry.Sources...)
} else {
collated[sn] = thisEntry
}
}
out := make([]*ServiceIntentionsConfigEntry, 0, len(collated))
for _, entry := range collated {
out = append(out, entry)
}
sort.Slice(out, func(i, j int) bool {
a := out[i]
b := out[j]
if a.PartitionOrDefault() < b.PartitionOrDefault() {
return true
} else if a.PartitionOrDefault() > b.PartitionOrDefault() {
return false
}
if a.NamespaceOrDefault() < b.NamespaceOrDefault() {
return true
} else if a.NamespaceOrDefault() > b.NamespaceOrDefault() {
return false
}
return a.Name < b.Name
})
connect: intentions are now managed as a new config entry kind "service-intentions" (#8834) - Upgrade the ConfigEntry.ListAll RPC to be kind-aware so that older copies of consul will not see new config entries it doesn't understand replicate down. - Add shim conversion code so that the old API/CLI method of interacting with intentions will continue to work so long as none of these are edited via config entry endpoints. Almost all of the read-only APIs will continue to function indefinitely. - Add new APIs that operate on individual intentions without IDs so that the UI doesn't need to implement CAS operations. - Add a new serf feature flag indicating support for intentions-as-config-entries. - The old line-item intentions way of interacting with the state store will transparently flip between the legacy memdb table and the config entry representations so that readers will never see a hiccup during migration where the results are incomplete. It uses a piece of system metadata to control the flip. - The primary datacenter will begin migrating intentions into config entries on startup once all servers in the datacenter are on a version of Consul with the intentions-as-config-entries feature flag. When it is complete the old state store representations will be cleared. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up. - The secondary datacenters continue to run the old intentions replicator until all servers in the secondary DC and primary DC support intentions-as-config-entries (via serf flag). Once this condition it met the old intentions replicator ceases. - The secondary datacenters replicate the new config entries as they are migrated in the primary. When they detect that the primary has zeroed it's old state store table it waits until all config entries up to that point are replicated and then zeroes its own copy of the old state store table. We also record a piece of system metadata indicating this has occurred. We use this metadata to skip ALL of this code the next time the leader starts up.
2020-10-06 18:24:05 +00:00
return out
}