Files
coredns/plugin/pkg/proxy/persistent.go

186 lines
4.8 KiB
Go
Raw Normal View History

package proxy
import (
"crypto/tls"
feat(forward): add source_address directive (#8011) * fix(dnssec): avoid caching empty signing results (#7996) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * save only Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * do #8008 Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix Address used if tcp Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix bad using of dialer type Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in gRPC transport (#8006) * core: Add full TSIG verification in gRPC transport This PR add full TSIG verification in gRPC using dns.TsigVerify() so invalid signatures and timestamps are correctly detected instead of only checking key presence. Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in QUIC transport (#8007) * core: Add full TSIG verification in QUIC transport This PR add full TSIG verification in QUIC using dns.TsigVerify() Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(test): deduplicate TSIG test helpers (#8009) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil sigs on sign error (#7999) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil from ParseKeyFile on error (#8000) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnsserver): allow view server blocks in any declaration order (#8001) When using the view plugin, filtered and unfiltered server blocks can share the same zone and port. The zone overlap validation rejected this configuration when the unfiltered block was not declared last, because filtered configs treated an already-registered zone as an error. Skip the 'already defined' check for configs that have filter functions, since they are expected to coexist with an unfiltered catch-all block on the same zone/port. Fixes #7733 Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(doh): use per-connection local address for PROXY protocol (#8005) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(transfer): batch AXFR records by message size instead of count (#8002) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(tls): use temp dir for keylog test path (#8010) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address option to source_address Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address also in readme Add test Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Resolve change request in pr Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix ci lint Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Improve doc on source_address routing needs Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove added timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * add use of source address also for health check query Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * remove untrailing newline from health_test.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix file format Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Update plugin/forward/setup_test.go Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove dead code in TestHealthLocalAddress Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix misspelling Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Try to set default timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix format in health.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> --------- Signed-off-by: Filippo <filippo.ferrazini@gmail.com> Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Co-authored-by: Yong Tang <yong.tang.github@outlook.com> Co-authored-by: Umut Polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Cedric Wang <wangzongqi@msn.com>
2026-07-10 02:38:22 +02:00
"net"
"net/http"
"sort"
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
"sync"
"time"
"github.com/miekg/dns"
)
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
// a persistConn holds the dns.Conn, its creation time, and the last used time.
type persistConn struct {
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
c *dns.Conn
created time.Time
used time.Time
}
// Transport hold the persistent cache.
type Transport struct {
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
avgDialTime int64 // kind of average time of dial time
conns [typeTotalCount][]*persistConn // Buckets for udp and tcp connections
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
expire time.Duration // After this duration an idle connection is expired.
maxAge time.Duration // After this duration a connection is closed regardless of activity; 0 means unlimited.
maxIdleConns int // Max idle connections per protocol type; 0 means unlimited.
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
addr string
tlsConfig *tls.Config
httpClient *http.Client
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
proxyName string
feat(forward): add source_address directive (#8011) * fix(dnssec): avoid caching empty signing results (#7996) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * save only Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * do #8008 Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix Address used if tcp Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix bad using of dialer type Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in gRPC transport (#8006) * core: Add full TSIG verification in gRPC transport This PR add full TSIG verification in gRPC using dns.TsigVerify() so invalid signatures and timestamps are correctly detected instead of only checking key presence. Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in QUIC transport (#8007) * core: Add full TSIG verification in QUIC transport This PR add full TSIG verification in QUIC using dns.TsigVerify() Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(test): deduplicate TSIG test helpers (#8009) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil sigs on sign error (#7999) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil from ParseKeyFile on error (#8000) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnsserver): allow view server blocks in any declaration order (#8001) When using the view plugin, filtered and unfiltered server blocks can share the same zone and port. The zone overlap validation rejected this configuration when the unfiltered block was not declared last, because filtered configs treated an already-registered zone as an error. Skip the 'already defined' check for configs that have filter functions, since they are expected to coexist with an unfiltered catch-all block on the same zone/port. Fixes #7733 Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(doh): use per-connection local address for PROXY protocol (#8005) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(transfer): batch AXFR records by message size instead of count (#8002) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(tls): use temp dir for keylog test path (#8010) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address option to source_address Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address also in readme Add test Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Resolve change request in pr Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix ci lint Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Improve doc on source_address routing needs Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove added timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * add use of source address also for health check query Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * remove untrailing newline from health_test.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix file format Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Update plugin/forward/setup_test.go Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove dead code in TestHealthLocalAddress Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix misspelling Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Try to set default timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix format in health.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> --------- Signed-off-by: Filippo <filippo.ferrazini@gmail.com> Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Co-authored-by: Yong Tang <yong.tang.github@outlook.com> Co-authored-by: Umut Polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Cedric Wang <wangzongqi@msn.com>
2026-07-10 02:38:22 +02:00
localAddress net.IP
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
mu sync.Mutex
stop chan struct{}
}
func newTransport(proxyName, addr string) *Transport {
t := &Transport{
avgDialTime: int64(maxDialTimeout / 2),
conns: [typeTotalCount][]*persistConn{},
expire: defaultExpire,
addr: addr,
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
stop: make(chan struct{}),
proxyName: proxyName,
}
return t
}
// connManager manages the persistent connection cache for UDP and TCP.
func (t *Transport) connManager() {
ticker := time.NewTicker(defaultExpire)
defer ticker.Stop()
for {
select {
case <-ticker.C:
t.cleanup(false)
case <-t.stop:
t.cleanup(true)
return
}
}
}
// closeConns closes connections.
func closeConns(conns []*persistConn) {
for _, pc := range conns {
pc.c.Close()
}
}
// cleanup removes connections from cache.
func (t *Transport) cleanup(all bool) {
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
var toClose []*persistConn
t.mu.Lock()
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
now := time.Now()
staleTime := now.Add(-t.expire)
// Pre-compute max-age deadline outside the loop to avoid repeated time.Now() calls.
var maxAgeDeadline time.Time
if t.maxAge > 0 {
maxAgeDeadline = now.Add(-t.maxAge)
}
for transtype, stack := range t.conns {
if len(stack) == 0 {
continue
}
if all {
t.conns[transtype] = nil
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
toClose = append(toClose, stack...)
continue
}
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
// When max-age is set, use a linear scan to evaluate both the idle-timeout
// (expire, based on last-used time) and the max-age (based on creation time).
if t.maxAge > 0 {
var alive []*persistConn
for _, pc := range stack {
if !pc.used.After(staleTime) || pc.created.Before(maxAgeDeadline) {
toClose = append(toClose, pc)
} else {
alive = append(alive, pc)
}
}
t.conns[transtype] = alive
continue
}
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
// Original expire-only path: connections are sorted by "used"; use binary search.
if stack[0].used.After(staleTime) {
continue
}
// connections in stack are sorted by "used"
good := sort.Search(len(stack), func(i int) bool {
return stack[i].used.After(staleTime)
})
t.conns[transtype] = stack[good:]
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
toClose = append(toClose, stack[:good]...)
}
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
t.mu.Unlock()
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
// Close connections after releasing lock
closeConns(toClose)
}
// Yield returns the connection to transport for reuse.
func (t *Transport) Yield(pc *persistConn) {
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
// Check if transport is stopped before acquiring lock
select {
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
case <-t.stop:
// If stopped, don't return to pool, just close
pc.c.Close()
return
default:
}
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
pc.used = time.Now() // update used time
t.mu.Lock()
defer t.mu.Unlock()
transtype := t.transportTypeFromConn(pc)
if t.maxIdleConns > 0 && len(t.conns[transtype]) >= t.maxIdleConns {
pc.c.Close()
return
}
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
t.conns[transtype] = append(t.conns[transtype], pc)
}
// Start starts the transport's connection manager.
func (t *Transport) Start() { go t.connManager() }
// Stop stops the transport's connection manager.
func (t *Transport) Stop() { close(t.stop) }
// SetExpire sets the connection expire time in transport.
func (t *Transport) SetExpire(expire time.Duration) { t.expire = expire }
plugin/forward: add max_age option to enforce an absolute connection lifetime (#7903) * plugin/pkg/proxy: add max_age for per-connection lifetime cap Introduce a max_age setting on Transport that closes connections based on creation time, independent of idle-timeout (expire). Background: PR #7790 changed the connection pool from LIFO to FIFO for source-port diversity. Under FIFO, every connection is cycled through the pool and its used timestamp is refreshed continuously. When request rate is high enough that pool_size / request_rate < expire, no connection ever becomes idle and expire never fires. This prevents CoreDNS from opening new connections to upstreams that scale out (e.g. new Kubernetes pods behind a ClusterIP service with conntrack pinning). max_age addresses this by enforcing an absolute upper bound on connection lifetime regardless of activity: - persistConn gains a created field set at dial time. - Transport gains maxAge (default 0 = unlimited, preserving existing behaviour). - Dial(): rejects cached connections whose creation age exceeds max_age. - cleanup(): when maxAge > 0, uses a linear scan over both idle-timeout and max-age predicates; when maxAge == 0, preserves the original binary-search path on used time (sorted by FIFO insertion order). - Both hot paths pre-compute the deadline outside any inner loop to avoid repeated time.Now() calls. Tests added: - TestMaxAgeExpireByCreation: connection with old created but fresh used must be rejected even though idle-timeout would pass. - TestMaxAgeFIFORotation: three FIFO-rotated connections (old created, fresh used) must all be rejected, confirming that continuous rotation cannot prevent max-age expiry. Signed-off-by: cangming <cangming@cangming.app> * plugin/forward: add max_age option Expose Transport.SetMaxAge through the forward plugin so operators can set an absolute upper bound on connection lifetime via the Corefile. Usage: forward . 1.2.3.4 { max_age 30s } Default is 0 (unlimited), which preserves existing behaviour. A positive value causes connections older than max_age to be closed and re-dialled on the next request, ensuring CoreDNS reconnects to newly scaled-out upstream pods even under sustained load where the idle timeout (expire) would never fire. If max_age is set, it must not be less than expire; the parser rejects this combination at startup with a clear error message. Signed-off-by: cangming <cangming@cangming.app> --------- Signed-off-by: cangming <cangming@cangming.app>
2026-03-10 02:50:03 +08:00
// SetMaxAge sets the maximum lifetime of a connection regardless of activity.
// A value of 0 (default) disables max-age and connections are only closed by expire (idle-timeout).
func (t *Transport) SetMaxAge(maxAge time.Duration) { t.maxAge = maxAge }
perf(proxy): use mutex-based connection pool (#7790) * perf(proxy): use mutex-based connection pool The proxy package (used for example by the forward plugin) utilized an actor model where a single connManager goroutine managed connection pooling via unbuffered channels (dial, yield, ret). This design serialized all connection acquisition and release operations through a single goroutine, creating a bottleneck under high concurrency. This was observable as a performance degradation when using a single upstream backend compared to multiple backends (which sharded the bottleneck). Changes: - Removed dial, yield, and ret channels from the Transport struct. - Removed the connManager goroutine's request processing loop. - Implemented Dial() and Yield() using a sync.Mutex to protect the connection slice, allowing for fast concurrent access without context switching. - Downgraded connManager to a simple background cleanup loop that only handles connection expiration on a ticker. - Updated plugin/pkg/proxy/connect.go to use direct method calls instead of channel sends. - Updated tests to reflect the removal of internal channels. Benchmarks show that this change eliminates the single-backend bottleneck. Now a single upstream backend performs on par with multiple backends, and overall throughput is improved. The implementation aligns with standard Go patterns for connection pooling (e.g., net/http.Transport). Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address PR review for persistent.go - Named mutex field instead of embedding, to not expose Lock() and Unlock() - Move stop check outside of lock in Yield() - Close() without a separate goroutine - Change stop channel to struct Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: address code review feedback for conn pool - Switch from LIFO to FIFO connection selection for source port diversity, reducing DNS cache poisoning risk (RFC 5452). - Remove "clear entire cache" optimization as it was LIFO-specific. FIFO naturally iterates and skips expired connections. - Remove all goroutines for closing connections; collect connections while holding lock, close synchronously after releasing lock. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * fix: remove unused error consts No longer utilised after refactoring the channel based approach. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * feat(forward): add max_idle_conns option Add configurable connection pool limit for the forward plugin via the max_idle_conns Corefile option. Changes: - Add SetMaxIdleConns to proxy - Add maxIdleConns field to Forward struct - Add max_idle_conns parsing in forward plugin setup - Apply setting to each proxy during configuration - Update forward plugin README with new option By default the value is 0 (unbounded). When set, excess connections returned to the pool are closed immediately rather than cached. Also add a yield related test. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore(proxy): simple Dial by closing conns inline Remove toClose slice collection to reduce complexity. Instead close expired connections directly while iterating. Reduces complexity with negligible lock-time impact. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> * chore: fewer explicit Unlock calls Cleaner and less chance of forgetting to unlock on new possible code paths. Signed-off-by: Ville Vesilehto <ville@vesilehto.fi> --------- Signed-off-by: Ville Vesilehto <ville@vesilehto.fi>
2026-01-14 03:49:46 +02:00
// SetMaxIdleConns sets the maximum idle connections per transport type.
// A value of 0 means unlimited (default).
func (t *Transport) SetMaxIdleConns(n int) { t.maxIdleConns = n }
// SetTLSConfig sets the TLS config in transport.
func (t *Transport) SetTLSConfig(cfg *tls.Config) { t.tlsConfig = cfg }
// GetTLSConfig returns the TLS config in transport.
func (t *Transport) GetTLSConfig() *tls.Config { return t.tlsConfig }
feat(forward): add source_address directive (#8011) * fix(dnssec): avoid caching empty signing results (#7996) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * save only Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * do #8008 Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix Address used if tcp Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix bad using of dialer type Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in gRPC transport (#8006) * core: Add full TSIG verification in gRPC transport This PR add full TSIG verification in gRPC using dns.TsigVerify() so invalid signatures and timestamps are correctly detected instead of only checking key presence. Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * core: Add full TSIG verification in QUIC transport (#8007) * core: Add full TSIG verification in QUIC transport This PR add full TSIG verification in QUIC using dns.TsigVerify() Signed-off-by: Yong Tang <yong.tang.github@outlook.com> * Fix Signed-off-by: Yong Tang <yong.tang.github@outlook.com> --------- Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(test): deduplicate TSIG test helpers (#8009) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil sigs on sign error (#7999) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnssec): return nil from ParseKeyFile on error (#8000) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(dnsserver): allow view server blocks in any declaration order (#8001) When using the view plugin, filtered and unfiltered server blocks can share the same zone and port. The zone overlap validation rejected this configuration when the unfiltered block was not declared last, because filtered configs treated an already-registered zone as an error. Skip the 'already defined' check for configs that have filter functions, since they are expected to coexist with an unfiltered catch-all block on the same zone/port. Fixes #7733 Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(doh): use per-connection local address for PROXY protocol (#8005) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(transfer): batch AXFR records by message size instead of count (#8002) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix(tls): use temp dir for keylog test path (#8010) Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address option to source_address Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Rename local_address also in readme Add test Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Resolve change request in pr Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix ci lint Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Improve doc on source_address routing needs Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove added timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * add use of source address also for health check query Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * remove untrailing newline from health_test.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix file format Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Update plugin/forward/setup_test.go Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Remove dead code in TestHealthLocalAddress Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Fix misspelling Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * Try to set default timeout Signed-off-by: Filippo <filippo.ferrazini@gmail.com> * fix format in health.go Signed-off-by: Filippo <filippo.ferrazini@gmail.com> --------- Signed-off-by: Filippo <filippo.ferrazini@gmail.com> Signed-off-by: Yong Tang <yong.tang.github@outlook.com> Signed-off-by: umut-polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Ville Vesilehto <ville@vesilehto.fi> Co-authored-by: Yong Tang <yong.tang.github@outlook.com> Co-authored-by: Umut Polat <52835619+umut-polat@users.noreply.github.com> Co-authored-by: Cedric Wang <wangzongqi@msn.com>
2026-07-10 02:38:22 +02:00
// SetLocalAddress sets the local address in transport.
func (t *Transport) SetLocalAddress(addr net.IP) {
t.localAddress = addr
}
const (
defaultExpire = 10 * time.Second
minDialTimeout = 1 * time.Second
maxDialTimeout = 30 * time.Second
)