Files
EvoBGP/internal/pipeline/collect_parallel.go
T
DenozordecandCursor f8e0520817 feat(httpapi): harden control plane and speed up prefix refresh
Очередь job получила lease и fencing, телеметрия и прогресс deploy сохраняются, ревизия пишет manifest входов. Обновление префиксов не переписывает неизменённый снимок, большие наборы идут через COPY, а устаревшая ревизия не попадает в deploy. Бандл ноды кэшируется, prune и постановка job больше не держат горячий путь.

Co-authored-by: Cursor <[email protected]>
2026-11-02 17:23:59 +07:00

353 lines
9.2 KiB
Go

package pipeline
import (
"context"
"errors"
"evobgp/internal/logging"
"fmt"
"net/http"
"os"
"strings"
"sync"
"time"
"evobgp/internal/store"
)
func prefixRowsForSource(rows []store.PrefixRow, sourceKey string) []store.PrefixRow {
if len(rows) == 0 {
return nil
}
var out []store.PrefixRow
for _, row := range rows {
if row.Source == sourceKey {
out = append(out, row)
}
}
return out
}
func prefixCommunityKey(row store.PrefixRow) string {
comm := ""
if row.CommunityID != nil {
comm = strings.TrimSpace(*row.CommunityID)
}
return row.Prefix + "\x00" + comm
}
func collectASPrefixRows(ctx context.Context, st store.Backend, hc *http.Client, tenantID string, mod *store.Module, list []*store.ASEntry, priorSnapshot []store.PrefixRow) ([]store.PrefixRow, error) {
moduleID := mod.ID
legacy := strings.TrimSpace(os.Getenv("EVOBGP_ASN_RESOLVE")) == "0"
if legacy {
var rows []store.PrefixRow
for _, e := range list {
if !store.ValidASN(e.ASN) {
continue
}
comm := e.CommunityID
if comm == nil && mod.DefaultCommunityID != nil {
c := *mod.DefaultCommunityID
comm = &c
}
rows = append(rows, store.PrefixRow{Prefix: MaterializedASPrefixKey(e.ASN), CommunityID: comm, Source: "as_entry"})
}
return rows, nil
}
type entryResult struct {
rows []store.PrefixRow
metaID string
asn int64
holder string
count int64
err error
}
var valid []*store.ASEntry
for _, e := range list {
if e != nil && store.ValidASN(e.ASN) {
valid = append(valid, e)
}
}
sem := make(chan struct{}, collectConcurrency())
results := make([]entryResult, len(valid))
var wg sync.WaitGroup
for i, e := range valid {
wg.Add(1)
go func(idx int, entry *store.ASEntry) {
defer wg.Done()
sem <- struct{}{}
defer func() { <-sem }()
comm := entry.CommunityID
if comm == nil && mod.DefaultCommunityID != nil {
c := *mod.DefaultCommunityID
comm = &c
}
pfxs, holder, err := resolveASNForEntry(ctx, st, hc, entry.ASN)
if err != nil {
if staleOnUpstreamError() {
if staleRows, staleHolder, ok := staleASNPrefixes(st, priorSnapshot, entry.ASN); ok {
logStaleUpstream("asn", fmt.Sprintf("AS%d: %v", entry.ASN, err))
src := fmt.Sprintf("as:%d", entry.ASN)
rows := append([]store.PrefixRow(nil), staleRows...)
for i := range rows {
rows[i].CommunityID = comm
rows[i].Source = src
}
results[idx] = entryResult{
rows: rows,
metaID: entry.ID,
asn: entry.ASN,
holder: staleHolder,
count: int64(len(rows)),
}
return
}
if pfxs2, holder2, ok := asnCacheExpired(st, entry.ASN); ok {
logStaleUpstream("asn", fmt.Sprintf("AS%d expired cache: %v", entry.ASN, err))
src := fmt.Sprintf("as:%d", entry.ASN)
var rows []store.PrefixRow
for _, pfx := range pfxs2 {
rows = append(rows, store.PrefixRow{Prefix: pfx.String(), CommunityID: comm, Source: src})
}
results[idx] = entryResult{
rows: rows,
metaID: entry.ID,
asn: entry.ASN,
holder: holder2,
count: int64(len(pfxs2)),
}
return
}
}
results[idx] = entryResult{err: fmt.Errorf("resolve AS%d: %w", entry.ASN, err)}
return
}
src := fmt.Sprintf("as:%d", entry.ASN)
var rows []store.PrefixRow
for _, pfx := range pfxs {
rows = append(rows, store.PrefixRow{Prefix: pfx.String(), CommunityID: comm, Source: src})
}
results[idx] = entryResult{
rows: rows,
metaID: entry.ID,
asn: entry.ASN,
holder: holder,
count: int64(len(pfxs)),
}
}(i, e)
}
wg.Wait()
seenPfx := make(map[string]struct{})
var out []store.PrefixRow
var metaUpdates []store.ASEntryResolveMetaUpdate
now := time.Now().UTC()
for _, r := range results {
if r.err != nil {
return nil, r.err
}
if r.metaID != "" {
metaUpdates = append(metaUpdates, store.ASEntryResolveMetaUpdate{
EntryID: r.metaID,
ASNName: r.holder,
PrefixCount: r.count,
})
}
for _, row := range r.rows {
k := prefixCommunityKey(row)
if _, ok := seenPfx[k]; ok {
continue
}
seenPfx[k] = struct{}{}
out = append(out, row)
}
}
if len(metaUpdates) > 0 {
if err := st.UpdateASEntryResolveMetaBatch(tenantID, moduleID, metaUpdates, now); err != nil {
return nil, fmt.Errorf("as entry meta batch: %w", err)
}
}
return out, nil
}
func collectCDNPrefixRows(ctx context.Context, st store.Backend, hc *http.Client, tenantID string, mod *store.Module, sources []*store.CDNSource, prior modulePrior) ([]store.PrefixRow, error) {
moduleID := mod.ID
now := time.Now().UTC()
var valid []*store.CDNSource
for _, s := range sources {
if s != nil {
valid = append(valid, s)
}
}
prior, skipRows, err := loadSkippedCDNRows(ctx, st, tenantID, moduleID, valid, prior, now)
if err != nil {
return nil, err
}
type srcResult struct {
rows []store.PrefixRow
err error
}
results := make([]srcResult, len(valid))
sem := make(chan struct{}, collectConcurrency())
var wg sync.WaitGroup
for i, src := range valid {
wg.Add(1)
go func(idx int, src *store.CDNSource) {
defer wg.Done()
sem <- struct{}{}
defer func() { <-sem }()
sourceKey := cdnSourceKey(src.ID)
if shouldSkipCDNSourceFetch(src, now) {
cached := prefixRowsForSource(prior.rows, sourceKey)
if !prior.provided && skipRows != nil {
cached = skipRows[sourceKey]
}
if len(cached) > 0 {
results[idx] = srcResult{rows: cached}
return
}
}
rows, err := fetchCDNSourceRows(ctx, st, hc, tenantID, moduleID, mod, src, prior, now)
if err != nil {
if staleOnUpstreamError() {
cached, ok, cacheErr := staleCDNPrefixes(ctx, st, tenantID, moduleID, prior, src.ID)
if cacheErr != nil {
results[idx] = srcResult{err: cacheErr}
return
}
if ok {
logStaleUpstream("cdn", fmt.Sprintf("source %s: %v", src.ID, err))
results[idx] = srcResult{rows: cached}
return
}
}
results[idx] = srcResult{err: err}
return
}
results[idx] = srcResult{rows: rows}
}(i, src)
}
wg.Wait()
var fetchedRows []store.PrefixRow
var skipped int
var fetchedIDs []string
allIDs := make([]string, 0, len(valid))
for _, src := range valid {
allIDs = append(allIDs, src.ID)
}
for i, r := range results {
if r.err != nil {
if errors.Is(r.err, errCDNSnapshotRead) {
return nil, r.err
}
if cdnPartialOK() {
logging.Default().Info(fmt.Sprintf("pipeline: CDN partial skip source error: %v", r.err))
skipped++
continue
}
return nil, r.err
}
fetchedIDs = append(fetchedIDs, valid[i].ID)
fetchedRows = append(fetchedRows, r.rows...)
}
if skipped > 0 && len(fetchedRows) == 0 && len(valid) > 0 {
return nil, fmt.Errorf("cdn: all %d source(s) failed (partial ok)", len(valid))
}
base := prior.rows
if skipped > 0 && len(base) == 0 && !prior.provided && st != nil {
if snap, ok, err := st.GetModulePrefixSnapshot(ctx, tenantID, moduleID); err != nil {
return nil, err
} else if ok && snap != nil {
base = snap.Prefixes
}
}
return mergeCDNSnapshotRows(base, fetchedRows, fetchedIDs, allIDs), nil
}
func collectDomainPrefixRows(ctx context.Context, st store.Backend, hc *http.Client, mod *store.Module, profiles []*store.DohProfile, policy string, entries []*store.DomainEntry, priorSnapshot []store.PrefixRow) ([]store.PrefixRow, error) {
var validDom []*store.DomainEntry
for _, e := range entries {
if e != nil {
validDom = append(validDom, e)
}
}
type domResult struct {
rows []store.PrefixRow
err error
}
results := make([]domResult, len(validDom))
sem := make(chan struct{}, collectConcurrency())
var wg sync.WaitGroup
for i, e := range validDom {
wg.Add(1)
go func(idx int, entry *store.DomainEntry) {
defer wg.Done()
sem <- struct{}{}
defer func() { <-sem }()
comm := entry.CommunityID
if comm == nil && mod.DefaultCommunityID != nil {
c := *mod.DefaultCommunityID
comm = &c
}
addrs, err := resolveDomainIPsCached(ctx, st, hc, profiles, policy, entry.FQDN)
if err != nil {
if staleOnUpstreamError() {
if cached, ok := staleDomainPrefixes(priorSnapshot, entry.FQDN); ok {
logStaleUpstream("domain", fmt.Sprintf("%q: %v", entry.FQDN, err))
rows := append([]store.PrefixRow(nil), cached...)
for i := range rows {
if rows[i].CommunityID == nil {
rows[i].CommunityID = comm
}
}
results[idx] = domResult{rows: rows}
return
}
}
results[idx] = domResult{err: fmt.Errorf("resolve domain %q: %w", entry.FQDN, err)}
return
}
src := "domain:" + strings.TrimSpace(entry.FQDN)
var rows []store.PrefixRow
for _, ip := range addrs {
cidr := ipToHostPrefix(ip)
if cidr == "" {
continue
}
rows = append(rows, store.PrefixRow{
Prefix: cidr,
CommunityID: comm,
Source: src,
})
}
results[idx] = domResult{rows: rows}
}(i, e)
}
wg.Wait()
seen := make(map[string]struct{})
var out []store.PrefixRow
for _, r := range results {
if r.err != nil {
return nil, r.err
}
for _, row := range r.rows {
key := row.Prefix + "|" + row.Source
if _, ok := seen[key]; ok {
continue
}
seen[key] = struct{}{}
out = append(out, row)
}
}
return out, nil
}