Files
seaweedfs/weed/server/filer_grpc_server_condition.go
T
Chris Lu 2dbf9c9108 s3: evaluate all conditional write headers under the object lock (#12211)
* filer_pb,filer: per-mutation preconditions on ObjectTransaction

A transaction's request-level condition evaluates one entry before any
mutation runs, so a commit guarding two paths — the object it publishes
and the upload directory it assembles — could not be expressed. Give each
ObjectMutation an optional condition (evaluated against condition_key,
else the mutation's target, under the same lock just before the mutation
applies) and report which mutation failed through failed_mutation on the
response, so callers can tell a mutation guard's rejection apart from the
request-level precondition's.

* s3: evaluate multipart completion conditions under the object lock

CompleteMultipartUpload checked conditional headers gateway-side before
committing, and the routed commit transaction guarded only the upload
directory — nothing made If-None-Match: * atomic with the publish, so
concurrent conditional completions of the same absent key could all
report success.

The request's reducible write condition (If-None-Match/If-Match, star or
single strong ETag) now rides the commit as the transaction's request
condition on the owner filer: on the object path for unversioned buckets,
falling back to the .versions directory for suspended and enabled
versioning. The upload-exists guard moves onto the PUT mutation, keeping
its check-before-publish ordering. Conditions the reducer cannot express
stay on the distributed write lock, where check and commit already
serialize.

* filer,s3: resolve the latest version first for versioned conditional writes

* filer: keep directory entries holding content out of the absent-slot rule

* s3: keep a gateway-side upload-exists check for filers without mutation conditions

* filer: honor null-version-is-latest resolving prefer-alt conditions

* drop scratch multipart-race repro programs

* s3: note the write lock does not cover routed writes on non-reducible conditions

* filer: mirror the gateway's remote-etag precedence in condition evaluation

storedEntryETag only knew the extended attribute and the chunk fallback,
so a synced remote entry compared chunk-derived ETags while the gateway
compares the remote ETag. Follow the same precedence: extended attribute,
remote ETag while the entry still holds the remote bytes, Md5, then the
chunk hash — and no ETag for an entry holding none.

* s3: reduce all conditional write headers into owner-lock clauses

Time-based headers, combined conditions, ETag lists, and weak ETags stayed
on the distributed write lock because buildWriteCondition could not express
them, leaving their gateway check racing owner-routed writes. Reduce every
parseable header form to WriteCondition clauses — multiple clauses compose
AND, ETag candidates carry the raw list for the filer's canonicalization,
and RFC 7232 precedence keeps If-Unmodified-Since hidden by If-Match and
If-Modified-Since hidden by If-None-Match — so the owner filer evaluates
them under the same object-path lock as the commit. Only unparseable
headers still fall back, where the lock path reports the same parse error.
2026-10-29 11:49:48 -07:00

189 lines
6.5 KiB
Go

package weed_server
import (
"fmt"
"strconv"
"strings"
"time"
"github.com/seaweedfs/seaweedfs/weed/filer"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
"google.golang.org/protobuf/proto"
)
// conditionIsSet reports whether a condition asks for any check at all.
func conditionIsSet(cond *filer_pb.WriteCondition) bool {
return cond != nil && len(cond.Clauses) > 0
}
// writeConditionSatisfied reports whether the precondition holds against the
// current entry (nil if absent), evaluated under the path lock. Every clause
// must hold (logical AND).
func writeConditionSatisfied(cond *filer_pb.WriteCondition, current *filer.Entry) bool {
for _, c := range cond.Clauses {
if !clauseSatisfied(c, current) {
return false
}
}
return true
}
// clauseSatisfied evaluates one primitive against the current entry. For the
// ETag kinds, etags is a set: IF_ETAG_MATCH holds when the current ETag equals
// any member, IF_ETAG_NOT_MATCH when it equals none. The IF_EXTENDED_* kinds are
// generic guards on an extended attribute used to enforce object-lock (legal
// hold and retention) without S3 knowledge in the filer.
func clauseSatisfied(c *filer_pb.WriteCondition_Clause, current *filer.Entry) bool {
exists := current != nil
switch c.Kind {
case filer_pb.WriteCondition_NONE:
return true
case filer_pb.WriteCondition_IF_NOT_EXISTS:
return !exists
case filer_pb.WriteCondition_IF_EXISTS:
return exists
case filer_pb.WriteCondition_IF_ETAG_MATCH:
return exists && etagInSet(storedEntryETag(current), c.Etags, c.AllowWeak)
case filer_pb.WriteCondition_IF_ETAG_NOT_MATCH:
return !exists || !etagInSet(storedEntryETag(current), c.Etags, c.AllowWeak)
case filer_pb.WriteCondition_IF_UNMODIFIED_SINCE:
return !exists || current.Attr.Mtime.Unix() <= c.UnixTime
case filer_pb.WriteCondition_IF_MODIFIED_SINCE:
return !exists || current.Attr.Mtime.Unix() > c.UnixTime
case filer_pb.WriteCondition_IF_EXTENDED_NOT_EQUAL:
if !exists {
return true
}
v, ok := current.Extended[c.ExtKey]
return !ok || string(v) != c.ExtValue
case filer_pb.WriteCondition_IF_EXTENDED_TIME_ELAPSED:
if !exists {
return true
}
// An optional gate scopes the guard: when gate_key is set, the time check
// only applies if extended[gate_key] == gate_value. This lets the gateway
// express governance bypass (enforce retention only for COMPLIANCE mode)
// without reading the entry — the filer decides under the lock.
if c.GateKey != "" {
gv, gok := current.Extended[c.GateKey]
if !gok || string(gv) != c.GateValue {
return true
}
}
v, ok := current.Extended[c.ExtKey]
if !ok {
return true
}
deadline, err := strconv.ParseInt(strings.TrimSpace(string(v)), 10, 64)
if err != nil {
// An unparseable retention deadline is treated as still in force, so
// a malformed attribute fails safe (write blocked) rather than open.
return false
}
return deadline <= time.Now().Unix()
case filer_pb.WriteCondition_IF_CHUNKS_EQUAL:
return chunkFidsEqual(current, c.Fids)
case filer_pb.WriteCondition_IF_ENTRY_EQUAL:
if !exists || c.ExpectedEntry == nil {
return !exists && c.ExpectedEntry == nil
}
// Compare both sides in serialized form on clones: chunks are matched
// by their fid only — the stored entry may carry the restored file_id
// while an expected one built from a metadata event does not.
expected := proto.Clone(c.ExpectedEntry).(*filer_pb.Entry)
filer_pb.BeforeEntrySerialization(expected.Chunks)
actual := proto.Clone(current.ToProtoEntry()).(*filer_pb.Entry)
filer_pb.BeforeEntrySerialization(actual.Chunks)
return proto.Equal(actual, filer.FromPbEntry("", expected).ToProtoEntry())
default:
// An unrecognized clause kind (e.g. from a newer client) must not be
// treated as satisfied, which would silently bypass the guard. Fail
// closed so the write is blocked rather than slipping through.
return false
}
}
// chunkFidsEqual compares the stored chunk fids (absent entry = none) against
// expected as multisets; chunk order carries no meaning for needle liveness.
func chunkFidsEqual(current *filer.Entry, expected []string) bool {
var chunks []*filer_pb.FileChunk
if current != nil {
chunks = current.GetChunks()
}
if len(chunks) != len(expected) {
return false
}
counts := make(map[string]int, len(expected))
for _, fid := range expected {
counts[fid]++
}
for _, chunk := range chunks {
fid := chunk.GetFileIdString()
if counts[fid] == 0 {
return false
}
counts[fid]--
}
return true
}
// etagInSet reports whether stored matches any candidate. A strong comparison
// (allowWeak false) treats a weak ETag as never equal; a weak comparison
// ignores the W/ marker on both sides.
func etagInSet(stored string, candidates []string, allowWeak bool) bool {
for _, c := range candidates {
if etagEqual(stored, c, allowWeak) {
return true
}
}
return false
}
func etagEqual(stored, expected string, allowWeak bool) bool {
sv, sWeak := canonicalETag(stored)
ev, eWeak := canonicalETag(expected)
// RFC 7232 strong comparison: a weak ETag on either side never matches.
if !allowWeak && (sWeak || eWeak) {
return false
}
return sv == ev
}
// canonicalETag splits off the weak (W/) marker before stripping quotes, so a
// weak ETag like W/"abc" yields ("abc", true).
func canonicalETag(etag string) (value string, weak bool) {
etag = strings.TrimSpace(etag)
if strings.HasPrefix(etag, "W/") {
return strings.Trim(etag[len("W/"):], `"`), true
}
return strings.Trim(etag, `"`), false
}
// storedEntryETag mirrors the S3 gateway's ETag precedence — the stored
// Seaweed ETag extended attribute, then the remote ETag while the entry still
// holds the remote bytes, then Md5, then the chunk hash (an empty object has
// no ETag) — so conditional comparisons match what the gateway computes,
// without coupling the filer to S3 request handling.
func storedEntryETag(entry *filer.Entry) string {
if v, ok := entry.Extended[s3_constants.ExtETagKey]; ok && len(v) > 0 {
return normalizeETag(string(v))
}
if remote := entry.Remote; remote != nil && remote.RemoteETag != "" &&
entry.FileSize == uint64(remote.RemoteSize) &&
(remote.LastLocalSyncTsNs > 0 || (len(entry.GetChunks()) == 0 && len(entry.Content) == 0)) {
return normalizeETag(remote.RemoteETag)
}
if len(entry.Md5) > 0 {
return fmt.Sprintf("%x", entry.Md5)
}
if len(entry.GetChunks()) == 0 {
return ""
}
return normalizeETag(filer.ETagChunks(entry.GetChunks()))
}
func normalizeETag(etag string) string {
return strings.Trim(etag, `"`)
}