mirror of
https://github.com/seaweedfs/seaweedfs.git
synced 2026-09-15 11:00:51 +02:00
Chris Lu
47b491b53c
* filer: stamp a log position on lookup and remote-cache responses Metadata events are logged after their store write and stamped with the filer clock. Reading that clock before serving an entry therefore gives a timestamp with a causal guarantee: every event at or below it is reflected in the returned entry. Clients caching filer state can use it as the entry's version to order the response against subscription events, including events committed before the call but delivered after it. * mount: version open file handles by filer log position A subscription event refreshing an open handle did a second lookup; a transient failure left the handle pinned to its old entry with no retry, since the subscription cursor had already advanced. The deeper problem is ordering: the handle is a cache written by three unordered channels — the async invalidation worker, local mutation acks, and open-time lookups — and overwriting cached state safely requires knowing which write is newer. The filer log timestamp is that order, and it now travels with every value instead of being derived out of band. Events carry it natively; lookup and remote-cache responses carry the log position stamped before the serving read; mutation acks carry it in their returned event; and the local store pairs each read with a version cursor advanced under the same lock as the store write. Each handle records the version its entry reflects, and one rule replaces the per-site reasoning: state at or below the handle's version is old news and must not be installed. The invalidation itself applies the event's own entry — no lookup, so no transient-failure window — except under a cached parent, where the store entry is the ordered merge of the event and anything applied since, and its version outranks the event's. An uncached parent receives no store writes, so a hit there would be a stale leftover masking the event. A vacated path (delete, rename away) keeps the last entry so unlinked-but-open reads still work. Directory builds version the completed directory at the listing snapshot and re-invalidate buffered events at that version, since their mid-build refresh ran against an incomplete store. The tests replay every race this replaces machinery for: rollback of a newer local flush (queued, cached, and read-through), stale leftovers under uncached parents, the build window including abort, handles opened after an event was queued, events landing mid-lookup, and undelivered events at remote-cache time across a filer failover. * filer: serialize the log position fence with mutations, stamp mutation acks The fence stamped before an unlocked entry read could precede state the read returned: a mutation writes storage first and assigns its event timestamp only at notify time, so a lookup racing that window handed the mount an entry newer than its fence, and the event's later delivery looked like fresh news — destroying dirty pages for a change the handle already had. The mutation handlers already hold an exclusive per-path lock across read, write, and notify; the lookup and remote-cache reads now take it shared around the stamp and the read, making the fence exact: everything at or below it is in the entry, nothing above it is. A no-change update returns success without an event, leaving the mount nothing to fence with even though the response confirms current state. Create and update acks now carry a log position stamped under the same lock, and the mount falls back to it whenever the ack has no event. Also regenerate the VT marshalers, which the earlier generation missed: without them a VT round-trip silently zeroed every log position. * java: sync filer.proto * mount: scope store versions to what they vouch for; atomic handle install The store's version cursor claimed too much. Advanced by local mutation acks and directory listing snapshots, it inflated the version of store reads for unrelated paths whose events the subscription still owed, and those events were then fenced out permanently. The cursor now tracks subscription progress only — events arrive in log order, so everything at or below it has been delivered for every path — and a completed listing records its snapshot as a per-directory floor instead of a global claim. Local acks never touch it: they version their own handle directly. Buffered build events advance the cursor at delivery, since their store write may never happen (abort) while their invalidation is already queued; their read-through directory pairs no store read with it, and rename fragments are applied first. Concurrent first opens raced: a slower opener's older lookup could overwrite the newer entry a faster opener had installed, while the monotonic version kept the newer timestamp — an old entry fenced at a new version, immune to every correcting event. Entry and version are now installed as one decision under the handle map lock, and an install that does not outrank the handle's version is dropped. The remote-cache commit also escaped the fence: it wrote storage and notified without the path lock, so a lookup's shared-locked fence and read could land between the two and hand out the cached state under-versioned. The commit now re-reads and writes under the exclusive path lock, and backs off entirely when the entry changed during the download — the concurrent writer supersedes the cached content. * mount: floors gate store applies; installs respect handle users; renames join the fence A directory floor certifies the listing state as of its snapshot, but a delayed event at or below the floor was still applied to the store — rolling the content back to pre-snapshot state while the floor kept claiming the snapshot version, so the correcting events were fenced out of every future read. Events are now gated against the affected directory's floor, each half of a rename independently. Fences are lower bounds: a listing or lookup can include a mutation whose event has not been delivered yet, and that event later passes every gate carrying state the handle already holds. Such a re-delivery now advances the version without destroying dirty pages or reinstalling the entry — invalidating local writes over a no-op was the real damage in every remaining under-fence window, including the unlocked listing snapshot, which no per-path lock can serialize. The concurrent-open install moved from the map lock to the handle lock every reader, writer, and invalidation synchronizes on, and rejects what cannot improve the handle: dirty state (local writes would be lost), unversioned lookup responses (they cannot outrank anything, and two zero-version opens must not overwrite each other), and anything not strictly newer. New handles are still fully initialized before the map exposes them. Renames committed metadata and emitted events with no path lock, so a lookup could read the renamed state under a fence preceding its events. Both rename handlers now hold the source and destination locks, ordered by path, across commit and notification; descendants of a renamed directory are not individually locked and rely on the no-op re-delivery handling above. * mount: per-entry store versions replace the cursor and directory floors The store's aggregate versions — a global subscription cursor and per-directory listing floors — were versions at coarser granularity than the values they described, and every over-claiming bug in this series traced to that gap: an aggregate vouching for state its source never saw. Each store entry now carries the filer log position of the write that produced it — the event that applied it, or the listing snapshot that inserted it, recorded in the store's key-value space under the same lock as the entry write. The store becomes what the handle already is: a last-writer-wins register with one rule, install only what outranks the current claim. The cursor, the floors, their advancement rules, the pairing ordering constraint, and the floor gating all collapse into that rule. Applies are gated per entry, each half of a rename independently; an unversioned local write clears the claim its content no longer proves; version records lingering after a bulk folder wipe cannot fence a recreate, since a claim only blocks while its entry exists. Listing inserts are stamped at build completion, before the buffered replay so newer replayed events override the stamp. Filer side, the fence dance every versioned read must perform is now a single choke point, fencedFindEntry, so a future read RPC gets the lock-serialized stamp by construction rather than by convention. * mount: judge no-op re-deliveries against an immutable base, not the live entry The equal-state skip compared the incoming event to the live handle entry, but local writes mutate the live entry — size, timestamps, chunks — so a delayed event re-delivering the base the handle was opened with no longer matched, and the installer destroyed the dirty pages and rolled the entry back over nothing new. The handle now keeps an immutable snapshot of the filer state it last installed or acknowledged, refreshed at every install and mutation ack (flush acks snapshot the request entry before the id mapping mutates it), and the no-op judgment runs against that base: an event carrying the base brings nothing, whatever the live entry has diverged to since. * mount: tombstones for versioned deletes, absence floors, copy enrollment Four gaps in the per-entry version protocol, all the same shape: a versioned fact with nothing carrying its version. A deletion is a fact about a path with no entry left to hold it — clearing the record let a delayed older event resurrect the deleted path, permanently, since the deletion's own redelivery is dedup-suppressed. Versioned deletes now leave a tombstone record that fences without an entry; renames tombstone their source the same way. Plain records still only block while their entry exists, so records lingering after a bulk folder wipe cannot fence a recreate. A completed listing proves absences as well as presences: a name it omitted was deleted as of the snapshot, and a delayed create below the snapshot re-creates it. The snapshot is kept per directory strictly as an absence fence, consulted only when a path has neither an entry nor a version record — present entries carry their own versions and never touch it, which is what separates this from the over-claiming floor it replaces. A rebuild against a pre-upgrade filer returns no snapshot; stamping now clears the children's records in that case, so a reinserted entry cannot reactivate the stale claim its previous incarnation left behind and reject valid events below it. Server-side copies installed the copied entry without enrolling in the base protocol, so the copy's own event differed from the stale pre-copy base and destroyed writes made to the destination after the copy. The install now refreshes the base and takes its version from the fenced readback. * mount: deletion facts outlive the cache's knowledge of the entry A versioned delete of a path the store held no entry for recorded nothing, so a delayed older event recreated the path — permanently, with the deletion's redelivery dedup-suppressed. The tombstone is now written whenever a versioned event vacates a path: the deletion is a fact about the path, not about what this cache happened to hold. For an absent entry, the listing's absence floor now speaks whatever older record remains: a tombstone at one position does not exhaust what is known about the path when a newer snapshot has confirmed the name still absent, and an event between the two was slipping past both. A committed copy whose readback failed installed a synthesized base with local timestamps; the copy's real event legitimately differs from it, and was read as foreign state — destroying writes made to the destination after the copy. The handle now marks that its own event is en route and adopts that event's state as the base without touching the live entry or the dirty pages; the adoption is one-shot, so a genuinely foreign event still invalidates. * mount: authoritative acks cancel pending event adoption; tombstones scoped and pruned The copy-event adoption flag could outlive its purpose: a flush after the failed readback installs a newer base and advances the version, the copy's own event is then version gated without consuming the flag, and the next genuinely foreign event was silently adopted — base advanced, live entry and dirty pages untouched — leaving the mount to later overwrite that remote change. Every local acknowledgment now installs its base through one helper that also cancels any pending adoption: the ack supersedes the mutation the adoption was waiting for. Tombstones were written for every versioned delete under the mount and survived directory eviction by design, growing LevelDB with historical deletions on delete-heavy mounts. They are now scoped to directories whose cached state the fence actually protects — an uncached parent never serves from the store nor applies the resurrecting insert — and a completed listing prunes the direct-child tombstones its absence floor supersedes, leaving only those above the snapshot. The store gains a key-prefix visitor for the sweep. * mount: acked saves install their value; trailer snapshots; direct-child prune range A version must never advance without its value. saveEntry stamped any open handle with the acknowledgment's version, but a handle opened while the save was in flight holds the pre-mutation entry — stamping it fenced out the events carrying the state it lacked, permanently, with the local apply performing no invalidation and the redelivery deduplicated. The acknowledged entry is now installed together with its version, through the same guarded install the racing-open path uses: under the handle lock, only when it outranks the handle, never over dirty local writes. Empty listings return no in-band snapshot — a snapshot-only response would be read as an entry by older consumers — so directories that end empty gained no absence floor and their tombstones were never pruned. The filer now sends the snapshot in the stream trailer, which older clients ignore, and the client reads it when no in-band snapshot arrived. Empty directories get real floors, their tombstones prune, and their buffered replays gain the snapshot filter instead of the replay-all fallback. Version records now encode the parent directory and name separated by a NUL, making a directory's direct children one contiguous key range: the tombstone prune scans exactly them under the cache lock, instead of walking every descendant record — the whole store, for root. * mount: fix dirty-page loss, uid/gid base, download race, copy adopt, leak; dedup Correctness fixes from the versioned-invalidation review: - A foreign delete/rename-away of a file held open with unflushed local writes destroyed the dirty pages unconditionally. A process may keep writing to an unlinked-but-open file and those writes were already acknowledged; preserve the pages when the handle is dirty. - downloadRemoteEntry stored the handle's base with filer-side uid/gid while every candidate it is later compared against is in local form, so under a non-identity UidGidMapper an unchanged re-delivery looked foreign and force-destroyed dirty pages. Map the base to local. - downloadRemoteEntry wrote the entry/base/version triple under only the handle's shared lock, so two concurrent reads of the same remote-only file could tear it. Serialize the install with a dedicated mutex (invalidation is already excluded by the exclusive handle lock). - A committed server-side copy whose readback failed adopted the FIRST event past the version gate as its base; a foreign write delivered first was silently swallowed. Adopt only an event whose content matches the synthesized base — the copy's own event — and install any other normally. - The deferred-create path relied on AcquireFileHandle installing the passed entry on a pre-existing handle, which the version rework dropped. Restore that install in the compat wrapper; the versioned open path keeps its gated install. Growth and hot-path cost: - Per-entry version records and tombstones leaked when a directory was evicted or read-through without a rebuild. An uncached directory gates its own inserts, so its records fence nothing; clear a directory's child version records when it is wiped for eviction. - FindEntry paid for the version KvGet on every lookup/getattr cache hit and threw it away. FindEntry now reads only the entry; the hot lookupEntry cache-hit path skips the version entirely. Cleanups: - Extract ackVersionTsNs over the shared response interface, replacing the metadata-event-else-log-ts snippet copy-pasted at four ack sites. - Extract acquireRenamePathLocks, replacing the verbatim sorted two-path lock fence in both rename handlers. * mount: no resurrection on foreign delete, version no-event acks, gate downloads, tighten copy adopt Follow-ups to the review patches: - Preserving dirty pages on a foreign delete let the next flush pass the isDeleted guard and CreateEntry, resurrecting the remotely-unlinked name. Mark the handle deleted in the vacate branch: the open fd can still read its buffered writes, but a flush no longer recreates the file. - A no-event acknowledgment (log fence only) synthesized a metadata event with TsNs 0, so the cache stored the entry unversioned and an older subscriber event rolled it back. Stamp the synthesized event with the ack's log position at all four ack sites. - downloadRemoteEntry serialized its install but did not check the version, so an older response arriving last overwrote the entry/base while the monotonic version kept the newer value, fencing corrections out. Install only when the response is at least as new as the handle. - sameEntryContent compared only size and chunks, so a foreign chmod with unchanged content was adopted as the copy's own event. Compare everything except server-assigned timestamps, so a metadata-only foreign change installs instead. * mount: trim comments to the non-obvious why The versioning work accumulated multi-line comment blocks restating what the code says. Keep the constraint a reader cannot derive — why a fence is exact, why a version must not advance without its value, why an uncached parent's records fence nothing — and drop the rest. * mount: distinguish rename from delete, tighten the download and adopt gates - A rename emits a nil old-path invalidation just like an unlink, so the vacate branch marked the handle deleted and later writes through the already-open descriptor were skipped instead of persisted. Carry the delete/rename distinction on the invalidation and mark only an actual delete. - The remote-download install accepted an unversioned response regardless of the handle's version, so during a rolling upgrade a delayed response could install stale content under a newer version. Require the response to be at least as new, with one exception: a handle still lacking local chunks takes the content anyway — it cannot read without it — but does not claim the response's log position. - Copy-event adoption returned without installing, so a foreign touch arriving before the copy's own event lost its timestamps. Content is unchanged either way, so the dirty pages stay valid; a clean handle now takes the entry, while a dirty one keeps its diverged version. * mount: one directory floor instead of a record per child; agree on TTL Review feedback: - Build completion wrote one KV record per direct child inside the cache write lock, so a large directory stalled every other cache operation for O(children) store writes. The directory's listing snapshot already covers every child it saw; make that floor the version for any child without a record of its own, and a child earns a record only when a later event touches it. One map write per build replaces the per-child writes, with the same fencing. - The presence probe read the store directly and so counted a TTL-expired entry as present, judging the path by a record describing content that has logically vanished. It now applies the same expiry the read path does, and an expired path falls back to its directory floor. - Preserve ErrNotFound identity when the commit-time re-read finds the object deleted, so callers still surface a 404. - Assert the rename-away source fence timestamp in the invalidation test. Also record the tombstone ceiling: distinct deleted names in a cached directory accumulate until it is rebuilt or evicted, which prunes everything at or below the new snapshot. * mount: pin the fence's clock domain instead of letting skew decide A log-position fence is stamped by one filer's clock under that filer's in-process lock, so comparing it to an event another filer logged is comparing two unrelated clocks. The two error directions are not equally costly: applying an event the fence already covered is a re-apply the base-equality check absorbs, while skipping one it does not cover leaves the handle holding exactly the state the event was meant to correct, with the subscription cursor already past it — the unhealable staleness this whole PR exists to remove. So refuse to guess. Fences now carry the signature of the filer that stamped them, and a handle records it alongside the position. An event is only fenced out when the filer that logged it is the one that stamped the fence — the logging filer appends its own signature, so its presence identifies the clock domain. Events from any other filer are applied. Positions taken from events keep comparing as before; the subscription already delivers those in order. The invalidation callback takes a struct now: it carries the path, entry, position, delete/rename distinction, and signatures, and was about to need a fifth positional parameter. * mount: follow a foreign rename; key page invalidation on content, not equality - A rename's old-path invalidation now carries the destination, and the handle follows the file there: an open fd tracks the inode, and leaving it on the old path made its next flush recreate that name instead of updating the renamed file. - Dirty pages overlay content, so only a content change invalidates them. Keying that on exact equality meant any timestamp-only event destroyed them, which the copy-adoption marker existed to paper over — a foreign touch could consume the marker and leave the copy's own event to drop the post-copy writes. Comparing content instead makes the marker unnecessary, so it is gone: a metadata-only event keeps the overlay, and a dirty handle keeps its diverged entry unless foreign content supersedes it. - A remote download response that is merely older is now refused even when the handle still lacks chunks; only an unversioned one is taken (and claims no position), since an older response's content predates what the handle reflects. - A refused or unversioned download no longer publishes to the metadata cache, where a zero-position event would clear the entry's version and let an older subscriber event roll the cache back. * mount: page invalidation keys on content alone; unversioned writes claim no position - sameEntryContent compared everything but timestamps, so a foreign chmod, chown, or xattr change counted as a content change and destroyed the dirty-page overlay. It was strict only to serve the copy-adoption marker, which is gone; its one caller now asks the question it actually needs — did the bytes change — so metadata-only events leave the overlay alone. - A rename over an existing file destroys that file, but its open handle was left live and still pointed at the name the renamed source now occupies, so its flush could overwrite it. MovePath already reports the displaced inode; mark that handle deleted. - An acknowledgment was refused whenever its position was numerically lower, even when a different filer stamped the fence it lost to. Two known, differing signatures mean unrelated clocks, so the comparison no longer applies there; unknown signatures still compare as before. - A local write with no log position behind it now records that explicitly instead of deleting its version record. Absence means the directory listing covers the path, which is why the snapshot floor applies; local content the listing never saw must not inherit it, or the events that would correct it are fenced out. * mount: widen the existing lookup functions instead of forking WithVersion twins The versioning work grew a parallel function for every accessor that needed to return a log position — lookupEntryWithVersion beside lookupEntry, maybeLoadEntryWithVersion beside maybeLoadEntry, FindEntryWithVersion beside FindEntry, AcquireFileHandleWithVersion beside AcquireFileHandle, advanceEntryVersion beside advanceEntryVersionTsNs, plus a getPbEntryWithVersion wrapper and an InsertListedEntriesForTest hook. Two names for one operation is two places to keep in step, and the split let callers pick the one that happened to compile. Each pair is now the single original name carrying the position, with callers that do not want it discarding it. filer_pb.GetEntry returns the fence its response already carried rather than a mount-side wrapper re-issuing the lookup, and InsertEntry takes the position its content reflects rather than a test-only twin that inserted without one. The one behavioural knot the merge exposed: AcquireFileHandle had been installing the entry on a pre-existing handle only in its unversioned form, which conflated 'the caller is authoritative' with 'the lookup had no version'. Deferred create is the only caller that means the former, so it now installs explicitly and the map function just acquires.
874 lines
31 KiB
Go
874 lines
31 KiB
Go
package router
|
|
|
|
import (
|
|
"context"
|
|
"sort"
|
|
"strconv"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/seaweedfs/seaweedfs/weed/glog"
|
|
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3_constants"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3lifecycle"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3lifecycle/engine"
|
|
"github.com/seaweedfs/seaweedfs/weed/s3api/s3lifecycle/reader"
|
|
)
|
|
|
|
// SiblingLister inspects the surviving versions of a versioned key.
|
|
// nil receiver or an error means "unknown" — callers suppress. Four
|
|
// queries: Survivors paginates the .versions/ container plus the bare
|
|
// null version (used by sole-survivor and bootstrap); LookupVersion
|
|
// fetches a single version file by id (used by pointer-transition
|
|
// routing to read the displaced version's identity and mtime);
|
|
// ListVersions paginates every version file in the .versions/
|
|
// container (used to compute NoncurrentIndex when a NewerNoncurrent
|
|
// rule is active); LookupNullVersion returns the bare-key entry that
|
|
// represents the null version (used by pointer-transition routing
|
|
// when oldID is empty and to include the null in expansion ranks).
|
|
type SiblingLister interface {
|
|
Survivors(ctx context.Context, bucket, objectKey string) (Survivors, error)
|
|
LookupVersion(ctx context.Context, bucket, objectKey, versionID string) (*filer_pb.Entry, error)
|
|
ListVersions(ctx context.Context, bucket, objectKey string) ([]*filer_pb.Entry, error)
|
|
LookupNullVersion(ctx context.Context, bucket, objectKey string) (entry *filer_pb.Entry, explicit bool, err error)
|
|
}
|
|
|
|
// Survivors describes the state under .versions/<key>/ plus the bare
|
|
// null-version that exists when versioning was turned on after the
|
|
// object was first PUT (s3api_object_versioning.go treats <bucket>/<key>
|
|
// as a regular file, the null version, in that case).
|
|
type Survivors struct {
|
|
Count int // entries under .versions/<key>/, capped at 2
|
|
LoneEntry *filer_pb.Entry // populated when Count == 1
|
|
HasNullVersion bool // bare <bucket>/<key> exists as a regular file
|
|
}
|
|
|
|
// Match is one (event, action) pair where EvaluateAction fired. The
|
|
// dispatcher runs `LifecycleDelete` at DueTime; identity-CAS in the RPC
|
|
// guards against drift between schedule time and dispatch time.
|
|
//
|
|
// For NoncurrentDays / NewerNoncurrent on a versioned bucket, ObjectKey
|
|
// is the seaweedfs storage path (logical-key + ".versions/" + version_id)
|
|
// so the dispatcher can locate the specific version, and VersionID
|
|
// carries the AWS-visible version ID separately.
|
|
type Match struct {
|
|
Key s3lifecycle.ActionKey
|
|
Action *engine.CompiledAction
|
|
Result s3lifecycle.EvalResult
|
|
EventTs time.Time
|
|
DueTime time.Time
|
|
Bucket string
|
|
ObjectKey string
|
|
VersionID string
|
|
Identity *EntryIdentity
|
|
}
|
|
|
|
// EntryIdentity is the schedule-time CAS witness; the dispatcher serializes
|
|
// it into the LifecycleDelete request. The fields mirror
|
|
// s3_lifecycle_pb.EntryIdentity but stay in-package so the router doesn't
|
|
// pull a proto dependency.
|
|
type EntryIdentity struct {
|
|
MtimeNs int64
|
|
Size int64
|
|
HeadFid string
|
|
ExtendedHash []byte
|
|
}
|
|
|
|
// Route returns the matches that fire for ev against snap. Only active
|
|
// event-driven actions are considered; SCAN_AT_DATE and DISABLED bypass
|
|
// this path.
|
|
func Route(ctx context.Context, snap *engine.Snapshot, ev *reader.Event, now time.Time, lister SiblingLister) []Match {
|
|
if snap == nil || ev == nil {
|
|
return nil
|
|
}
|
|
keys := snap.BucketActionKeys(ev.Bucket)
|
|
if len(keys) == 0 {
|
|
return nil
|
|
}
|
|
// Bootstrap-expanded version event: sibling state is pre-computed,
|
|
// info.Key is the LOGICAL key so rule prefixes match. Skip the
|
|
// meta-log path's version-folder skip.
|
|
if ev.BootstrapVersion != nil {
|
|
return routeBootstrapVersion(snap, ev, keys)
|
|
}
|
|
|
|
versioned := snap.BucketVersioned(ev.Bucket)
|
|
|
|
// .versions/ directory metadata update: when ExtLatestVersionIdKey
|
|
// changes, the OLD pointer value names a version that's now
|
|
// noncurrent. Route NoncurrentDays / NewerNoncurrent for it without
|
|
// waiting for the next bootstrap.
|
|
if versioned && ev.NewEntry != nil && ev.OldEntry != nil && ev.NewEntry.IsDirectory && isVersionsContainerKey(ev.Key) {
|
|
if !hasActiveEventDrivenAction(snap, keys, s3lifecycle.ActionKindNoncurrentDays) &&
|
|
!hasActiveEventDrivenAction(snap, keys, s3lifecycle.ActionKindNewerNoncurrent) {
|
|
return nil
|
|
}
|
|
return routePointerTransition(ctx, snap, ev, keys, lister)
|
|
}
|
|
|
|
// EXP_DM can fire on two version-folder events: the marker create
|
|
// (sole survivor immediately) and a noncurrent hard-delete that
|
|
// leaves only the marker behind. Both reach routeSoleSurvivorMarker.
|
|
if versioned && isVersionFolderPath(ev.Key) {
|
|
if !hasActiveEventDrivenAction(snap, keys, s3lifecycle.ActionKindExpiredDeleteMarker) {
|
|
return nil
|
|
}
|
|
return routeSoleSurvivorMarker(ctx, snap, ev, keys, lister)
|
|
}
|
|
|
|
if ev.NewEntry == nil {
|
|
return nil
|
|
}
|
|
info := buildObjectInfo(ev, versioned)
|
|
if info == nil {
|
|
return nil
|
|
}
|
|
eventTime := time.Unix(0, ev.TsNs)
|
|
|
|
var matches []Match
|
|
for _, key := range keys {
|
|
action := snap.Action(key)
|
|
if action == nil || !action.IsActive() {
|
|
continue
|
|
}
|
|
if action.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
// (kind, info) shape gate: ABORT_MPU only fires on MPU init events,
|
|
// and other kinds never do. Without this an MPU init would be
|
|
// matched against NONCURRENT_DAYS (IsLatest=false reads as a
|
|
// non-current version) and the dispatcher would BLOCK on empty
|
|
// version_id.
|
|
if info.IsMPUInit && key.ActionKind != s3lifecycle.ActionKindAbortMPU {
|
|
continue
|
|
}
|
|
if !info.IsMPUInit && key.ActionKind == s3lifecycle.ActionKindAbortMPU {
|
|
continue
|
|
}
|
|
// Schedule from the per-kind due moment. ExpirationDate is
|
|
// rule-relative (the date IS the moment); other kinds are
|
|
// ModTime-relative. Using ModTime+Delay for ExpirationDate
|
|
// (Delay=0) puts dueTime at the entry's mtime — a backdated
|
|
// object's mtime is BEFORE the rule's date, so the eligibility
|
|
// check below would skip it. ComputeDueAt encapsulates both
|
|
// shapes; the dispatcher's identity-CAS catches drift if the
|
|
// object changes meanwhile.
|
|
dueTime := s3lifecycle.ComputeDueAt(action.Rule, key.ActionKind, info)
|
|
if dueTime.IsZero() {
|
|
continue
|
|
}
|
|
res := s3lifecycle.EvaluateAction(action.Rule, key.ActionKind, info, dueTime)
|
|
if res.Action == s3lifecycle.ActionNone {
|
|
continue
|
|
}
|
|
matches = append(matches, Match{
|
|
Key: key,
|
|
Action: action,
|
|
Result: res,
|
|
EventTs: eventTime,
|
|
DueTime: dueTime,
|
|
Bucket: ev.Bucket,
|
|
ObjectKey: ev.Key,
|
|
Identity: buildIdentity(ev),
|
|
})
|
|
}
|
|
return matches
|
|
}
|
|
|
|
// routePointerTransition handles a .versions/ container update where
|
|
// ExtLatestVersionIdKey changed: the OLD pointer value names a version
|
|
// that just became noncurrent. Two lookup shapes:
|
|
//
|
|
// - Pure NoncurrentVersionExpirationDays without NewerNoncurrentVersions:
|
|
// a single LookupVersion of oldID is enough — the displaced version
|
|
// is the only one that newly entered eligibility for this rule.
|
|
//
|
|
// - Any active NewerNoncurrentVersions rule: a pointer flip shifts
|
|
// every prior noncurrent's rank by one, so the version that *just
|
|
// crossed* the keep-count threshold needs evaluation too. List the
|
|
// full .versions/ container, rank newest-first, and route every
|
|
// eligible noncurrent. Identity-CAS handles dedup with earlier
|
|
// schedules.
|
|
//
|
|
// Without this branch the worker has to wait for the next bootstrap to
|
|
// schedule retention on a freshly-noncurrent version.
|
|
func routePointerTransition(ctx context.Context, snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey, lister SiblingLister) []Match {
|
|
if lister == nil {
|
|
return nil
|
|
}
|
|
logical := strings.TrimSuffix(ev.Key, s3_constants.VersionsFolder)
|
|
if logical == "" {
|
|
return nil
|
|
}
|
|
oldID := string(ev.OldEntry.Extended[s3_constants.ExtLatestVersionIdKey])
|
|
newID := string(ev.NewEntry.Extended[s3_constants.ExtLatestVersionIdKey])
|
|
if oldID == newID {
|
|
// Same id means the update didn't transition the pointer.
|
|
return nil
|
|
}
|
|
// oldID == "" doesn't mean "nothing displaced": a bare null may
|
|
// have been the implicit/explicit latest before the pointer
|
|
// flipped to a real id.
|
|
// newID == "" means a suspended-versioning write cleared the
|
|
// pointer and made the bare null current. The cached
|
|
// ExtLatestVersionMtimeKey may still hold the prior latest's
|
|
// mtime (stale), so we must NOT use successorModTimeFromContainer
|
|
// in that case — derive the successor clock from the null entry's
|
|
// mtime instead. latestIDForExpand carries the same substitution
|
|
// so the expansion path's latestPos lookup matches the null sibling.
|
|
var successor time.Time
|
|
latestIDForExpand := newID
|
|
if newID == "" {
|
|
nullEntry, _, err := lister.LookupNullVersion(ctx, ev.Bucket, logical)
|
|
if err != nil {
|
|
glog.V(2).Infof("lifecycle router: lookup null %s/%s: %v", ev.Bucket, logical, err)
|
|
return nil
|
|
}
|
|
if nullEntry == nil || nullEntry.Attributes == nil {
|
|
return nil
|
|
}
|
|
successor = time.Unix(nullEntry.Attributes.Mtime, int64(nullEntry.Attributes.MtimeNs))
|
|
latestIDForExpand = "null"
|
|
} else {
|
|
successor = successorModTimeFromContainer(ev.NewEntry)
|
|
}
|
|
if successor.IsZero() {
|
|
return nil
|
|
}
|
|
if needsFullExpansion(snap, keys) {
|
|
return routePointerTransitionExpand(ctx, snap, ev, keys, lister, logical, latestIDForExpand, successor)
|
|
}
|
|
return routePointerTransitionDisplaced(ctx, snap, ev, keys, lister, logical, oldID, successor)
|
|
}
|
|
|
|
// needsFullExpansion reports whether any active event-driven rule on
|
|
// this bucket cares about NoncurrentIndex (NewerNoncurrentVersions > 0
|
|
// in either NoncurrentDays or pure-count NewerNoncurrent).
|
|
func needsFullExpansion(snap *engine.Snapshot, keys []s3lifecycle.ActionKey) bool {
|
|
for _, k := range keys {
|
|
if k.ActionKind != s3lifecycle.ActionKindNoncurrentDays && k.ActionKind != s3lifecycle.ActionKindNewerNoncurrent {
|
|
continue
|
|
}
|
|
a := snap.Action(k)
|
|
if a == nil || !a.IsActive() || a.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
if a.Rule != nil && a.Rule.NewerNoncurrentVersions > 0 {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// successorModTimeFromContainer reads the cached latest-version mtime
|
|
// from the .versions/ container's Extended map.
|
|
// updateLatestVersionInDirectory writes it via setCachedListMetadata
|
|
// alongside ExtLatestVersionIdKey, but the directory's own
|
|
// Attributes.Mtime is preserved across pointer updates — using it
|
|
// directly would let a stale dir mtime trigger expiration immediately.
|
|
// Returns zero time if the cached mtime is missing or unparseable; the
|
|
// caller suppresses in that case.
|
|
func successorModTimeFromContainer(entry *filer_pb.Entry) time.Time {
|
|
raw, ok := entry.Extended[s3_constants.ExtLatestVersionMtimeKey]
|
|
if !ok || len(raw) == 0 {
|
|
return time.Time{}
|
|
}
|
|
secs, err := strconv.ParseInt(string(raw), 10, 64)
|
|
if err != nil || secs <= 0 {
|
|
return time.Time{}
|
|
}
|
|
return time.Unix(secs, 0)
|
|
}
|
|
|
|
// routePointerTransitionDisplaced is the single-lookup path: only the
|
|
// displaced version's noncurrent eligibility could have changed, so
|
|
// fetching just its file is enough. oldID == "" routes the bare null
|
|
// version instead — it was the implicit latest before the pointer
|
|
// flipped to a real id.
|
|
func routePointerTransitionDisplaced(ctx context.Context, snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey, lister SiblingLister, logical, oldID string, successor time.Time) []Match {
|
|
var displaced *filer_pb.Entry
|
|
displacedID := oldID
|
|
if oldID == "" {
|
|
nullEntry, _, err := lister.LookupNullVersion(ctx, ev.Bucket, logical)
|
|
if err != nil {
|
|
glog.V(2).Infof("lifecycle router: lookup null version %s/%s: %v", ev.Bucket, logical, err)
|
|
return nil
|
|
}
|
|
if nullEntry == nil {
|
|
return nil
|
|
}
|
|
displaced = nullEntry
|
|
displacedID = "null"
|
|
} else {
|
|
entry, err := lister.LookupVersion(ctx, ev.Bucket, logical, oldID)
|
|
if err != nil {
|
|
glog.V(2).Infof("lifecycle router: lookup displaced version %s/%s/%s: %v", ev.Bucket, logical, oldID, err)
|
|
return nil
|
|
}
|
|
displaced = entry
|
|
}
|
|
if displaced == nil || displaced.Attributes == nil {
|
|
return nil
|
|
}
|
|
// Prefer the explicit demotion stamp on the displaced entry over the
|
|
// container-derived successor. Stamp is written by the S3 PUT handler
|
|
// at the moment the pointer flipped; container value is derived from
|
|
// the new latest's mtime and may drift across retries.
|
|
effectiveSuccessor := successor
|
|
if stamp := s3lifecycle.SuccessorFromEntryStamp(displaced); !stamp.IsZero() {
|
|
effectiveSuccessor = stamp
|
|
}
|
|
idx := 0
|
|
info := &s3lifecycle.ObjectInfo{
|
|
Key: logical,
|
|
ModTime: time.Unix(displaced.Attributes.Mtime, int64(displaced.Attributes.MtimeNs)),
|
|
Size: int64(displaced.Attributes.FileSize),
|
|
IsLatest: false,
|
|
IsDeleteMarker: string(displaced.Extended[s3_constants.ExtDeleteMarkerKey]) == "true",
|
|
NoncurrentIndex: &idx,
|
|
SuccessorModTime: effectiveSuccessor,
|
|
}
|
|
if tags := extractTags(displaced.Extended); len(tags) > 0 {
|
|
info.Tags = tags
|
|
}
|
|
return emitNoncurrentMatches(snap, ev, keys, info, displaced, displacedID, effectiveSuccessor)
|
|
}
|
|
|
|
// routePointerTransitionExpand routes only the versions that newly
|
|
// became eligible by the pointer flip:
|
|
//
|
|
// - rank 0: the displaced version (newly noncurrent), needed for the
|
|
// pure-NoncurrentDays clock,
|
|
// - rank == rule.NewerNoncurrentVersions for each active rule that
|
|
// gates on count: the version at exactly that rank just crossed
|
|
// from kept to expired.
|
|
//
|
|
// Emitting every eligible noncurrent on every PUT would push
|
|
// O(versions) heap entries per flip — Schedule.Add doesn't dedup, so
|
|
// identity-CAS at dispatch only stops the wasted RPC, not the heap
|
|
// growth. Bootstrap still owns full backfill.
|
|
func routePointerTransitionExpand(ctx context.Context, snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey, lister SiblingLister, logical, newID string, successor time.Time) []Match {
|
|
rawVersions, err := lister.ListVersions(ctx, ev.Bucket, logical)
|
|
if err != nil {
|
|
glog.V(2).Infof("lifecycle router: list versions %s/%s: %v", ev.Bucket, logical, err)
|
|
return nil
|
|
}
|
|
// Include the bare null version in the sibling set: count-based
|
|
// ranks are wrong if a pre-versioning or suspended-null entry
|
|
// exists outside .versions/. ID is "null" for sort + emit.
|
|
nullEntry, _, nullErr := lister.LookupNullVersion(ctx, ev.Bucket, logical)
|
|
if nullErr != nil {
|
|
glog.V(2).Infof("lifecycle router: lookup null %s/%s: %v", ev.Bucket, logical, nullErr)
|
|
return nil
|
|
}
|
|
type sibling struct {
|
|
entry *filer_pb.Entry
|
|
id string
|
|
}
|
|
siblings := make([]sibling, 0, len(rawVersions)+1)
|
|
for _, v := range rawVersions {
|
|
if v == nil || v.Attributes == nil {
|
|
continue
|
|
}
|
|
id := string(v.Extended[s3_constants.ExtVersionIdKey])
|
|
if id == "" {
|
|
continue
|
|
}
|
|
siblings = append(siblings, sibling{entry: v, id: id})
|
|
}
|
|
if nullEntry != nil && nullEntry.Attributes != nil {
|
|
siblings = append(siblings, sibling{entry: nullEntry, id: "null"})
|
|
}
|
|
if len(siblings) == 0 {
|
|
return nil
|
|
}
|
|
sort.SliceStable(siblings, func(i, j int) bool {
|
|
mi := siblings[i].entry.Attributes.Mtime*int64(1e9) + int64(siblings[i].entry.Attributes.MtimeNs)
|
|
mj := siblings[j].entry.Attributes.Mtime*int64(1e9) + int64(siblings[j].entry.Attributes.MtimeNs)
|
|
if mi != mj {
|
|
return mi > mj
|
|
}
|
|
return s3lifecycle.CompareVersionIds(siblings[i].id, siblings[j].id) < 0
|
|
})
|
|
// Resolve latestPos by finding newID. Default to -1 so a missing
|
|
// newID (race with the listing, or torn write) suppresses the
|
|
// expansion: we'd otherwise call the actual newest sibling "latest"
|
|
// against the pointer's intent and misrank every noncurrent.
|
|
// Bootstrap repairs state on the next walk.
|
|
latestPos := -1
|
|
if newID != "" {
|
|
for i, s := range siblings {
|
|
if s.id == newID {
|
|
latestPos = i
|
|
break
|
|
}
|
|
}
|
|
}
|
|
if latestPos < 0 {
|
|
glog.V(2).Infof("lifecycle router: pointer transition %s/%s: new id %s not found in listing", ev.Bucket, logical, newID)
|
|
return nil
|
|
}
|
|
noncurrentCount := len(siblings) - 1
|
|
|
|
// Collect the target noncurrent ranks: 0 (the freshly displaced)
|
|
// plus N for each active count-gated rule.
|
|
rankSet := map[int]struct{}{0: {}}
|
|
for _, k := range keys {
|
|
if k.ActionKind != s3lifecycle.ActionKindNoncurrentDays && k.ActionKind != s3lifecycle.ActionKindNewerNoncurrent {
|
|
continue
|
|
}
|
|
a := snap.Action(k)
|
|
if a == nil || !a.IsActive() || a.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
if a.Rule != nil && a.Rule.NewerNoncurrentVersions > 0 {
|
|
rankSet[a.Rule.NewerNoncurrentVersions] = struct{}{}
|
|
}
|
|
}
|
|
ranks := make([]int, 0, len(rankSet))
|
|
for r := range rankSet {
|
|
ranks = append(ranks, r)
|
|
}
|
|
sort.Ints(ranks)
|
|
|
|
var matches []Match
|
|
for _, rank := range ranks {
|
|
if rank >= noncurrentCount {
|
|
continue
|
|
}
|
|
// Convert noncurrent rank to position in the sorted slice,
|
|
// skipping the latest's slot.
|
|
i := rank
|
|
if rank >= latestPos {
|
|
i = rank + 1
|
|
}
|
|
s := siblings[i]
|
|
// Successor mtime: the entry directly newer than this one in
|
|
// the sorted list. When the next-newer slot is the latest,
|
|
// use the cached successor (the new latest's mtime); otherwise
|
|
// the immediate predecessor's mtime.
|
|
var thisSuccessor time.Time
|
|
if i > 0 && i-1 != latestPos {
|
|
thisSuccessor = time.Unix(siblings[i-1].entry.Attributes.Mtime, int64(siblings[i-1].entry.Attributes.MtimeNs))
|
|
} else {
|
|
thisSuccessor = successor
|
|
}
|
|
// Override with the explicit demotion stamp when present —
|
|
// PUT-time wall clock beats derived sibling mtime for accuracy
|
|
// and is immune to mtime edits on the sibling itself.
|
|
if stamp := s3lifecycle.SuccessorFromEntryStamp(s.entry); !stamp.IsZero() {
|
|
thisSuccessor = stamp
|
|
}
|
|
idx := rank
|
|
info := &s3lifecycle.ObjectInfo{
|
|
Key: logical,
|
|
ModTime: time.Unix(s.entry.Attributes.Mtime, int64(s.entry.Attributes.MtimeNs)),
|
|
Size: int64(s.entry.Attributes.FileSize),
|
|
IsLatest: false,
|
|
IsDeleteMarker: string(s.entry.Extended[s3_constants.ExtDeleteMarkerKey]) == "true",
|
|
NoncurrentIndex: &idx,
|
|
SuccessorModTime: thisSuccessor,
|
|
NumVersions: len(siblings),
|
|
}
|
|
if tags := extractTags(s.entry.Extended); len(tags) > 0 {
|
|
info.Tags = tags
|
|
}
|
|
matches = append(matches, emitNoncurrentMatches(snap, ev, keys, info, s.entry, s.id, thisSuccessor)...)
|
|
}
|
|
return matches
|
|
}
|
|
|
|
// emitNoncurrentMatches walks NoncurrentDays / NewerNoncurrent action
|
|
// keys and emits Matches for each one that fires. Shared between the
|
|
// single-lookup and full-expansion paths.
|
|
func emitNoncurrentMatches(snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey, info *s3lifecycle.ObjectInfo, entry *filer_pb.Entry, versionID string, successor time.Time) []Match {
|
|
eventTime := time.Unix(0, ev.TsNs)
|
|
identity := buildIdentityFromEntry(entry)
|
|
var matches []Match
|
|
for _, key := range keys {
|
|
if key.ActionKind != s3lifecycle.ActionKindNoncurrentDays && key.ActionKind != s3lifecycle.ActionKindNewerNoncurrent {
|
|
continue
|
|
}
|
|
action := snap.Action(key)
|
|
if action == nil || !action.IsActive() || action.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
clock := successor
|
|
if clock.IsZero() {
|
|
clock = info.ModTime
|
|
}
|
|
dueTime := clock.Add(action.Delay)
|
|
res := s3lifecycle.EvaluateAction(action.Rule, key.ActionKind, info, dueTime)
|
|
if res.Action == s3lifecycle.ActionNone {
|
|
continue
|
|
}
|
|
matches = append(matches, Match{
|
|
Key: key,
|
|
Action: action,
|
|
Result: res,
|
|
EventTs: eventTime,
|
|
DueTime: dueTime,
|
|
Bucket: ev.Bucket,
|
|
ObjectKey: info.Key,
|
|
VersionID: versionID,
|
|
Identity: identity,
|
|
})
|
|
}
|
|
return matches
|
|
}
|
|
|
|
// routeSoleSurvivorMarker emits an EXP_DM Match against the LOGICAL key
|
|
// (so the dispatcher can call deleteSpecificObjectVersion) with the
|
|
// marker's version_id. Handles two events: a marker create (the new
|
|
// entry IS the marker) and a noncurrent hard-delete that leaves the
|
|
// marker behind (the listing's lone entry IS the marker). The server
|
|
// re-checks before deleting.
|
|
func routeSoleSurvivorMarker(ctx context.Context, snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey, lister SiblingLister) []Match {
|
|
if lister == nil {
|
|
return nil
|
|
}
|
|
// Skip the listing RPC for events that can't possibly produce a
|
|
// sole-survivor marker: a regular non-marker version create/update
|
|
// always lands at Count >= 2.
|
|
if ev.NewEntry != nil && !isDeleteMarkerEntry(ev.NewEntry) {
|
|
return nil
|
|
}
|
|
logicalKey, ok := logicalKeyFromVersionPath(ev.Key)
|
|
if !ok {
|
|
return nil
|
|
}
|
|
s, err := lister.Survivors(ctx, ev.Bucket, logicalKey)
|
|
if err != nil {
|
|
glog.V(2).Infof("lifecycle router: survivors %s/%s: %v", ev.Bucket, logicalKey, err)
|
|
return nil
|
|
}
|
|
// Pre-versioning bare-key objects (the "null" version) live outside
|
|
// .versions/. Treating count==1 as sole-survivor while a null
|
|
// version exists would let lifecycle delete the marker and re-expose
|
|
// the old object.
|
|
if s.Count != 1 || s.HasNullVersion || s.LoneEntry == nil {
|
|
return nil
|
|
}
|
|
if !isDeleteMarkerEntry(s.LoneEntry) {
|
|
return nil
|
|
}
|
|
versionID := string(s.LoneEntry.Extended[s3_constants.ExtVersionIdKey])
|
|
if versionID == "" {
|
|
// Empty version_id would BLOCK at dispatch and freeze the cursor.
|
|
return nil
|
|
}
|
|
entry := s.LoneEntry
|
|
info := &s3lifecycle.ObjectInfo{
|
|
Key: logicalKey,
|
|
ModTime: time.Unix(entry.Attributes.Mtime, int64(entry.Attributes.MtimeNs)),
|
|
Size: int64(entry.Attributes.FileSize),
|
|
IsLatest: true,
|
|
IsDeleteMarker: true,
|
|
NumVersions: 1,
|
|
}
|
|
if tags := extractTags(entry.Extended); len(tags) > 0 {
|
|
info.Tags = tags
|
|
}
|
|
eventTime := time.Unix(0, ev.TsNs)
|
|
identity := buildIdentityFromEntry(entry)
|
|
var matches []Match
|
|
for _, key := range keys {
|
|
if key.ActionKind != s3lifecycle.ActionKindExpiredDeleteMarker {
|
|
continue
|
|
}
|
|
action := snap.Action(key)
|
|
if action == nil || !action.IsActive() || action.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
dueTime := info.ModTime.Add(action.Delay)
|
|
res := s3lifecycle.EvaluateAction(action.Rule, key.ActionKind, info, dueTime)
|
|
if res.Action == s3lifecycle.ActionNone {
|
|
continue
|
|
}
|
|
matches = append(matches, Match{
|
|
Key: key,
|
|
Action: action,
|
|
Result: res,
|
|
EventTs: eventTime,
|
|
DueTime: dueTime,
|
|
Bucket: ev.Bucket,
|
|
ObjectKey: logicalKey,
|
|
VersionID: versionID,
|
|
Identity: identity,
|
|
})
|
|
}
|
|
return matches
|
|
}
|
|
|
|
// routeBootstrapVersion handles a synthesized event from BucketBootstrapper.
|
|
// The bootstrap walker has already listed .versions/<key>/, sorted siblings
|
|
// newest-first, and stamped each one's IsLatest / NoncurrentIndex /
|
|
// SuccessorModTime. The router only needs to assemble ObjectInfo and run
|
|
// the match loop with the standard kind gates. ev.NewEntry is the version
|
|
// file itself; ev.Key is the version-folder path; the LOGICAL key from
|
|
// BootstrapVersion drives prefix matching and the dispatcher.
|
|
func routeBootstrapVersion(snap *engine.Snapshot, ev *reader.Event, keys []s3lifecycle.ActionKey) []Match {
|
|
bv := ev.BootstrapVersion
|
|
entry := ev.NewEntry
|
|
if entry == nil || entry.Attributes == nil || bv.LogicalKey == "" {
|
|
return nil
|
|
}
|
|
idx := bv.NoncurrentIndex
|
|
info := &s3lifecycle.ObjectInfo{
|
|
Key: bv.LogicalKey,
|
|
ModTime: time.Unix(entry.Attributes.Mtime, int64(entry.Attributes.MtimeNs)),
|
|
Size: int64(entry.Attributes.FileSize),
|
|
IsLatest: bv.IsLatest,
|
|
IsDeleteMarker: bv.IsDeleteMarker,
|
|
NumVersions: bv.NumVersions,
|
|
SuccessorModTime: bv.SuccessorModTime,
|
|
}
|
|
if !bv.IsLatest {
|
|
info.NoncurrentIndex = &idx
|
|
}
|
|
if tags := extractTags(entry.Extended); len(tags) > 0 {
|
|
info.Tags = tags
|
|
}
|
|
eventTime := time.Unix(0, ev.TsNs)
|
|
identity := buildIdentityFromEntry(entry)
|
|
var matches []Match
|
|
for _, key := range keys {
|
|
action := snap.Action(key)
|
|
if action == nil || !action.IsActive() || action.Mode != engine.ModeEventDriven {
|
|
continue
|
|
}
|
|
// ABORT_MPU never applies to a versioned object.
|
|
if key.ActionKind == s3lifecycle.ActionKindAbortMPU {
|
|
continue
|
|
}
|
|
// Noncurrent rules clock from when this version was replaced
|
|
// (SuccessorModTime), not from when it was originally written.
|
|
// Bootstrap populates SuccessorModTime; fall back to ModTime
|
|
// for the latest version (no successor exists).
|
|
clock := info.ModTime
|
|
if !info.IsLatest && !info.SuccessorModTime.IsZero() {
|
|
clock = info.SuccessorModTime
|
|
}
|
|
dueTime := clock.Add(action.Delay)
|
|
res := s3lifecycle.EvaluateAction(action.Rule, key.ActionKind, info, dueTime)
|
|
if res.Action == s3lifecycle.ActionNone {
|
|
continue
|
|
}
|
|
// Pin the version_id only for kinds the dispatcher needs to
|
|
// target by version: noncurrent retention and the marker
|
|
// itself. EXPIRATION_DAYS / EXPIRATION_DATE on the latest
|
|
// must NOT carry it — between schedule and dispatch a fresh
|
|
// PUT can land, and identity-CAS against the original
|
|
// version's bytes would still pass even though the latest has
|
|
// moved on. Empty VersionID makes the dispatcher fetch the
|
|
// current latest, where identity-CAS resolves to STALE_IDENTITY
|
|
// and bootstrap re-schedules with the new latest's identity.
|
|
var matchVersionID string
|
|
switch key.ActionKind {
|
|
case s3lifecycle.ActionKindNoncurrentDays,
|
|
s3lifecycle.ActionKindNewerNoncurrent,
|
|
s3lifecycle.ActionKindExpiredDeleteMarker:
|
|
matchVersionID = bv.VersionID
|
|
}
|
|
matches = append(matches, Match{
|
|
Key: key,
|
|
Action: action,
|
|
Result: res,
|
|
EventTs: eventTime,
|
|
DueTime: dueTime,
|
|
Bucket: ev.Bucket,
|
|
ObjectKey: bv.LogicalKey,
|
|
VersionID: matchVersionID,
|
|
Identity: identity,
|
|
})
|
|
}
|
|
return matches
|
|
}
|
|
|
|
// logicalKeyFromVersionPath extracts <logical> from <logical>.versions/<file>.
|
|
// Returns false for a bucket-root marker (path = ".versions/<file>"), since
|
|
// AWS S3 has no concept of an object at the bucket key itself.
|
|
func logicalKeyFromVersionPath(versionPath string) (string, bool) {
|
|
lastSlash := strings.LastIndex(versionPath, "/")
|
|
if lastSlash <= 0 {
|
|
return "", false
|
|
}
|
|
parent := versionPath[:lastSlash]
|
|
if !strings.HasSuffix(parent, s3_constants.VersionsFolder) {
|
|
return "", false
|
|
}
|
|
logical := strings.TrimSuffix(parent, s3_constants.VersionsFolder)
|
|
if logical == "" {
|
|
return "", false
|
|
}
|
|
return logical, true
|
|
}
|
|
|
|
// buildObjectInfo derives an ObjectInfo from a meta-log event. Returns
|
|
// nil for shapes the router can't classify safely: missing attributes,
|
|
// non-MPU directories, version-folder files (those route through
|
|
// routeSoleSurvivorMarker upstream when EXP_DM applies). On a versioned
|
|
// bucket the latest pointer lives in the .versions/ directory's
|
|
// Extended map; without it we leave NumVersions=0 so the bootstrap walk
|
|
// drives noncurrent retention.
|
|
func buildObjectInfo(ev *reader.Event, versioned bool) *s3lifecycle.ObjectInfo {
|
|
entry := ev.NewEntry
|
|
if entry == nil || entry.Attributes == nil {
|
|
return nil
|
|
}
|
|
if destKey, ok := mpuInitInfo(ev, entry); ok {
|
|
return &s3lifecycle.ObjectInfo{
|
|
Key: destKey,
|
|
ModTime: time.Unix(entry.Attributes.Mtime, int64(entry.Attributes.MtimeNs)),
|
|
IsMPUInit: true,
|
|
}
|
|
}
|
|
if entry.IsDirectory {
|
|
return nil
|
|
}
|
|
info := &s3lifecycle.ObjectInfo{
|
|
Key: ev.Key,
|
|
ModTime: time.Unix(entry.Attributes.Mtime, int64(entry.Attributes.MtimeNs)),
|
|
Size: int64(entry.Attributes.FileSize),
|
|
IsLatest: true,
|
|
NumVersions: 1,
|
|
}
|
|
if versioned {
|
|
if isVersionFolderPath(ev.Key) {
|
|
return nil
|
|
}
|
|
info.NumVersions = 0
|
|
}
|
|
if tags := extractTags(entry.Extended); len(tags) > 0 {
|
|
info.Tags = tags
|
|
}
|
|
if isDeleteMarkerEntry(entry) {
|
|
info.IsDeleteMarker = true
|
|
}
|
|
return info
|
|
}
|
|
|
|
// isVersionsContainerKey reports whether the bucket-relative key IS a
|
|
// .versions/ container itself (e.g. "logs/foo.versions"), as opposed to
|
|
// a file inside one. Used to recognize directory-level events whose
|
|
// Extended map carries the latest pointer for an object.
|
|
func isVersionsContainerKey(key string) bool {
|
|
if key == s3_constants.VersionsFolder {
|
|
// Bucket-root .versions: no logical object key.
|
|
return false
|
|
}
|
|
return strings.HasSuffix(key, s3_constants.VersionsFolder)
|
|
}
|
|
|
|
// isVersionFolderPath reports whether the bucket-relative key sits inside a
|
|
// .versions/ folder — i.e. the path's parent segment ends with the
|
|
// VersionsFolder suffix. Used by the versioned-bucket gate so the router
|
|
// skips version-file events that need sibling state to be classified
|
|
// safely.
|
|
func isVersionFolderPath(key string) bool {
|
|
idx := strings.LastIndex(key, "/")
|
|
if idx <= 0 {
|
|
return false
|
|
}
|
|
parent := key[:idx]
|
|
parentIdx := strings.LastIndex(parent, "/")
|
|
leaf := parent[parentIdx+1:]
|
|
return strings.HasSuffix(leaf, s3_constants.VersionsFolder)
|
|
}
|
|
|
|
// mpuInitInfo recognizes a multipart-upload init: a directory entry at
|
|
// `.uploads/<upload_id>` carrying the destination key in Extended. Sub-events
|
|
// for part uploads (deeper paths under the upload directory) are deliberately
|
|
// rejected — they ride a different mtime and would over-fire ABORT_MPU.
|
|
func mpuInitInfo(ev *reader.Event, entry *filer_pb.Entry) (destKey string, ok bool) {
|
|
uploadsPrefix := s3_constants.MultipartUploadsFolder + "/"
|
|
if !entry.IsDirectory || !strings.HasPrefix(ev.Key, uploadsPrefix) {
|
|
return "", false
|
|
}
|
|
rest := ev.Key[len(uploadsPrefix):]
|
|
if rest == "" || strings.ContainsRune(rest, '/') {
|
|
// `.uploads/` itself or `.uploads/<id>/<part>...`; not the init.
|
|
return "", false
|
|
}
|
|
keyBytes, hasKey := entry.Extended[s3_constants.ExtMultipartObjectKey]
|
|
if !hasKey || len(keyBytes) == 0 {
|
|
return "", false
|
|
}
|
|
return string(keyBytes), true
|
|
}
|
|
|
|
// buildIdentity captures the entry's schedule-time fingerprint for the CAS
|
|
// witness. Returns nil if the event has no entry to fingerprint (deletes).
|
|
func buildIdentity(ev *reader.Event) *EntryIdentity {
|
|
return buildIdentityFromEntry(ev.NewEntry)
|
|
}
|
|
|
|
func buildIdentityFromEntry(entry *filer_pb.Entry) *EntryIdentity {
|
|
if entry == nil {
|
|
return nil
|
|
}
|
|
id := &EntryIdentity{}
|
|
if entry.Attributes != nil {
|
|
// Mirror the server-side computeEntryIdentity encoding: Mtime
|
|
// (seconds) and MtimeNs (nanosecond component) combine into
|
|
// EntryIdentity.MtimeNs as nanoseconds-since-epoch.
|
|
id.MtimeNs = entry.Attributes.Mtime*int64(1e9) + int64(entry.Attributes.MtimeNs)
|
|
id.Size = int64(entry.Attributes.FileSize)
|
|
}
|
|
if len(entry.GetChunks()) > 0 {
|
|
// Meta-log events arrive with chunk.FileId cleared by
|
|
// BeforeEntrySerialization; GetFileIdString reconstructs it from
|
|
// Fid so the worker matches the server-side fingerprint.
|
|
id.HeadFid = entry.GetChunks()[0].GetFileIdString()
|
|
}
|
|
id.ExtendedHash = s3lifecycle.HashExtended(entry.Extended)
|
|
return id
|
|
}
|
|
|
|
func extractTags(ext map[string][]byte) map[string]string {
|
|
if len(ext) == 0 {
|
|
return nil
|
|
}
|
|
prefix := s3_constants.AmzObjectTagging + "-"
|
|
var out map[string]string
|
|
for k, v := range ext {
|
|
if !strings.HasPrefix(k, prefix) {
|
|
continue
|
|
}
|
|
if out == nil {
|
|
out = map[string]string{}
|
|
}
|
|
out[k[len(prefix):]] = string(v)
|
|
}
|
|
return out
|
|
}
|
|
|
|
// hasActiveEventDrivenAction gates I/O (e.g. sibling listing) on whether
|
|
// a match could actually fire. Mirrors the per-key filter in Route so
|
|
// disabled or scan-only actions don't pay the RPC.
|
|
func hasActiveEventDrivenAction(snap *engine.Snapshot, keys []s3lifecycle.ActionKey, kind s3lifecycle.ActionKind) bool {
|
|
for _, k := range keys {
|
|
if k.ActionKind != kind {
|
|
continue
|
|
}
|
|
a := snap.Action(k)
|
|
if a == nil {
|
|
continue
|
|
}
|
|
if a.IsActive() && a.Mode == engine.ModeEventDriven {
|
|
return true
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// isDeleteMarkerEntry mirrors every read site for ExtDeleteMarkerKey:
|
|
// production writes []byte("true").
|
|
func isDeleteMarkerEntry(entry *filer_pb.Entry) bool {
|
|
if entry == nil || len(entry.Extended) == 0 {
|
|
return false
|
|
}
|
|
v, ok := entry.Extended[s3_constants.ExtDeleteMarkerKey]
|
|
return ok && string(v) == "true"
|
|
}
|