Commit Graph
14 Commits
Author SHA1 Message Date
Chris LuandGitHub 746ed82662 remote.meta.sync: sync directories and remove files deleted from remote (#10184)
* remote.meta.sync: materialize directory entries, including empty ones

Pull metadata by walking the remote tree one directory level at a time
with a delimiter, so subdirectories, including empty ones, are listed as
their own entries and created locally. The previous flat listing only
returned files, so empty remote directories never appeared locally and
non-empty ones only existed as filer-synthesized parents.

* remote.meta.sync: remove local metadata for entries deleted from remote

After reconciling each directory, drop local entries whose remote source
is gone: files are deleted outright, and a directory removed from the
remote is descended into so its remote-backed children are cleaned while
local-only entries are kept. remote.meta.sync exposes -delete (default
on) and remote.mount.buckets reconciles the same way; a plain
remote.mount stays additive.

* remote.meta.sync: reconcile type swaps and prune emptied directories

- when the remote swaps an entry's type (file <-> directory), drop the
  stale local entry and recreate it with the right type; local-only
  entries are left alone
- mark synced directories remote-backed and clean a directory removed
  from the remote locally, deleting it once it holds no local-only
  entries, instead of re-listing the missing remote path
- treat a differing remote size or mtime, not only a newer mtime, as a
  change worth pulling
2026-07-01 12:14:19 -07:00
0e570d6a8f feat(remote.mount): add -metadataStrategy flag to control metadata caching (#8568)
* feat(remote): add -noSync flag to skip upfront metadata pull on mount

Made-with: Cursor

* refactor(remote): split mount setup from metadata sync

Extract ensureMountDirectory for create/validate; call pullMetadata
directly when sync is needed. Caller controls sync step for -noSync.

Made-with: Cursor

* fix(remote): validate mount root when -noSync so bad bucket/creds fail fast

When -noSync is used, perform a cheap remote check (ListBuckets and
verify bucket exists) instead of skipping all remote I/O. Invalid
buckets or credentials now fail at mount time.

Made-with: Cursor

* test(remote): add TestRemoteMountNoSync for -noSync mount and persisted mapping

Made-with: Cursor

* test(remote): assert no upfront metadata after -noSync mount

After remote.mount -noSync, run fs.ls on the mount dir and assert empty
listing so the test fails if pullMetadata was invoked eagerly.

Made-with: Cursor

* fix(remote): propagate non-ErrNotFound lookup errors in ensureMountDirectory

Return lookupErr immediately for any LookupDirectoryEntry failure that
is not filer_pb.ErrNotFound, so only the not-found case creates the
entry and other lookup failures are reported to the caller.

Made-with: Cursor

* fix(remote): use errors.Is for ErrNotFound in ensureMountDirectory

Replace fragile strings.Contains(lookupErr.Error(), ...) with
errors.Is(lookupErr, filer_pb.ErrNotFound) before calling CreateEntry.

Made-with: Cursor

* fix(remote): use LookupEntry so ErrNotFound is recognised after gRPC

Raw gRPC LookupDirectoryEntry returns a status error, not the sentinel,
so errors.Is(lookupErr, filer_pb.ErrNotFound) was always false. Use
filer_pb.LookupEntry which normalises not-found to ErrNotFound so the
mount directory is created when missing.

Made-with: Cursor

* test(remote): ignore weed shell banner in TestRemoteMountNoSync fs.ls count

Exclude master/filer and prompt lines from entry count so the assertion
checks only actual fs.ls output for empty -noSync mount.

Made-with: Cursor

* fix(remote.mount): use 0755 for mount dir, document bucket-less early return

Made-with: Cursor

* feat(remote.mount): replace -noSync with -metadataStrategy=lazy|eager

- Add -metadataStrategy flag (eager default, lazy skips upfront metadata pull)
- Accept lazy/eager case-insensitively; reject invalid values with clear error
- Rename TestRemoteMountNoSync to TestRemoteMountMetadataStrategyLazy
- Add TestRemoteMountMetadataStrategyEager and TestRemoteMountMetadataStrategyInvalid

Made-with: Cursor

* fix(remote.mount): validate strategy and remote before creating mount directory

Move strategy validation and validateMountRoot (lazy path) before
ensureMountDirectory so that invalid strategies or bad bucket/credentials
fail without leaving orphaned directory entries in the filer.

* refactor(remote.mount): remove unused remote param from ensureMountDirectory

The remote *RemoteStorageLocation parameter was left over from the old
syncMetadata signature. Only remoteConf.Name is used inside the function.

* doc(remote.mount): add TODO for HeadBucket-style validation

validateMountRoot currently lists all buckets to verify one exists.
Note the need for a targeted BucketExists method in the interface.

* refactor(remote.mount): use MetadataStrategy type and constants

Replace raw string comparisons with a MetadataStrategy type and
MetadataStrategyEager/MetadataStrategyLazy constants for clarity
and compile-time safety.

* refactor(remote.mount): rename MetadataStrategy to MetadataCacheStrategy

More precisely describes the purpose: controlling how metadata is
cached from the remote, not metadata handling in general.

* fix(remote.mount): remove validateMountRoot from lazy path

Lazy mount's purpose is to skip remote I/O. Validating via ListBuckets
contradicts that, especially on accounts with many buckets. Invalid
buckets or credentials will surface on first lazy access instead.

* fix(test): handle shell exit 0 in TestRemoteMountMetadataStrategyInvalid

The weed shell process exits with code 0 even when individual commands
fail — errors appear in stdout. Check output instead of requiring a
non-nil error.

* test(remote.mount): remove metadataStrategy shell integration tests

These tests only verify string output from a shell process that always
exits 0 — they cannot meaningfully validate eager vs lazy behavior
without a real remote backend.

---------

Co-authored-by: Chris Lu <chris.lu@gmail.com>
2026-03-12 15:21:07 -07:00
9012069bd7 chore: execute goimports to format the code (#7983)
* chore: execute goimports to format the code

Signed-off-by: promalert <promalert@outlook.com>

* goimports -w .

---------

Signed-off-by: promalert <promalert@outlook.com>
Co-authored-by: Chris Lu <chris.lu@gmail.com>
2026-01-07 13:06:08 -08:00
Chris LuandGitHub 69553e5ba6 convert error fromating to %w everywhere (#6995) 2025-07-16 23:39:27 -07:00
chrislu ec30a504ba refactor 2024-09-29 10:38:22 -07:00
chrislu 701abbb9df add IsResourceHeavy() to command interface 2024-09-28 20:23:01 -07:00
Shaun AlexanderandGitHub 766290411b filer.remote.gateway not filer.remote.sync for remote.mount.buckets help string. (#5934) 2024-08-23 20:41:33 -07:00
chrislu 26dbc6c905 move to https://github.com/seaweedfs/seaweedfs 2022-07-29 00:17:28 -07:00
chrislu 6793bc853c help message when in simulation mode 2022-05-31 14:48:46 -07:00
Chris Lu a8b0f8864d fix help message 2021-11-07 12:38:35 -08:00
Chris Lu 8e4b43a017 minor 2021-09-06 15:13:38 -07:00
Chris Lu 32e94de86a default auto trimming suffix to true 2021-09-05 13:27:35 -07:00
Chris Lu 60573fd3e2 option to map remote bucket to trimmed bucket name 2021-09-05 11:55:52 -07:00
Chris Lu d57d4c5f8f shell: add remote.mount.buckets 2021-09-04 21:37:25 -07:00