Chris Lu 9e98ec4b2e Share decoded metadata-log entries across subscriber replays (#9903) ...

perf(filer): share decoded log entries across metadata replays

Concurrent SubscribeMetadata replays of the same persisted log history each
opened a chunk reader per source filer and re-decoded the same files, so a
reconnect storm multiplied into many GB of buffers. Cache the decoded entries
of completed log files in a bounded LRU, coalescing concurrent loads with
single-flight and bounding concurrent decodes. Each hit is validated against
the file's current chunk set, so a file that received a late append is
reloaded rather than served stale; reads that stop on an unreachable chunk are
delivered but not cached so a transient outage re-probes on the next replay.

2026-06-09 13:34:11 -07:00

..

admin

feat(admin): export full cluster volume list as JSON (#9876)

2026-06-08 15:01:02 -07:00

cluster

s3: route object reads to the key's owner filer (#9806)

2026-06-03 00:12:28 -07:00

command

fix(filer.sync): replicate a rename as an atomic move, not a no-op update (#9895)

2026-06-09 12:54:28 -07:00

credential

grpc: don't tear down the shared master connection on a caller's own timeout (#9775)

2026-06-01 15:11:02 -07:00

filer

Share decoded metadata-log entries across subscriber replays (#9903)

2026-06-09 13:34:11 -07:00

filer_client

…

glog

…

iam

fix(iam): return a valid user ARN from CreateUser and GetUser (#9794)

2026-06-02 22:01:57 -07:00

iamapi

fix(iam): return a valid user ARN from CreateUser and GetUser (#9794)

2026-06-02 22:01:57 -07:00

images

…

kms

…

mount

Fix filer metadata-replay OOM under mount reconnect storms (#9901)

2026-06-09 11:43:12 -07:00

mq

fix(mq): don't cache topic non-existence on transient filer errors

2026-06-08 12:04:48 -07:00

notification

…

operation

operation: index VidCache by map instead of slice (#9853)

2026-06-07 11:46:57 -07:00

pb

Fix filer metadata-replay OOM under mount reconnect storms (#9901)

2026-06-09 11:43:12 -07:00

plugin/worker

Treat co-located volume servers as one fault domain when balancing and allocating (#9854)

2026-06-07 14:14:45 -07:00

query

…

remote_storage

fix(remote): correct content and permissions when syncing/caching remote objects (#9879)

2026-06-08 13:55:53 -07:00

replication

fix(filer.sync): replicate a rename as an atomic move, not a no-op update (#9895)

2026-06-09 12:54:28 -07:00

s3api

fix(s3api): clear stale object lock years (#9890)

2026-06-09 00:48:38 -07:00

security

security: hot-reload JWT signing keys on SIGHUP (#9826)

2026-06-04 22:26:08 -07:00

sequence

…

server

Fix filer metadata-replay OOM under mount reconnect storms (#9901)

2026-06-09 11:43:12 -07:00

sftpd

Don't mangle filer paths with the OS separator on Windows (#9878)

2026-06-08 13:56:02 -07:00

shell

fix(shell): move files into existing destination directories (#9887)

2026-06-08 23:42:13 -07:00

static

…

stats

feat(filer): object size distribution metric and dashboard panels (#9902)

2026-06-09 10:41:11 -07:00

storage

fix(seaweed-volume): stop EC shard deletion from phantom .dat on restart (#9874)

2026-06-08 22:10:16 -07:00

telemetry

…

topology

vacuum: compact a read-only volume when an explicit volumeId is given (#9861)

2026-06-07 22:42:51 -07:00

util

fix(http): accept no content delete responses (#9893)

2026-06-09 11:45:14 -07:00

wdclient

grpc: don't tear down the shared master connection on a caller's own timeout (#9775)

2026-06-01 15:11:02 -07:00

worker

ec.balance: verify shard landed on destination before deleting the source (#9858)

2026-06-07 21:31:53 -07:00

Makefile

…

weed.go

…