Merge branch 'upstream_master' into store_s3cred

# Conflicts:
#	weed/s3api/filer_util.go

.github/workflows/release.yml

@@ -3,8 +3,6 @@ name: Release
 on:
   push:
     branches: [ master ]
-  pull_request:
-    branches: [ master ]
 
 jobs:
 
@@ -16,14 +14,28 @@ jobs:
     - name: Check out code into the Go module directory
       uses: actions/checkout@v2
 
-#    - name: Go Release Binaries
+    - name: Go Release Binaries
-#      uses: wangyoucao577/go-release-action@v1.8
+      uses: wangyoucao577/go-release-action@v1.10
-#      with:
+      with:
-#        github_token: ${{ secrets.GITHUB_TOKEN }}
+        github_token: ${{ secrets.GITHUB_TOKEN }}
-#        goos: linux # default is 
+        goos: linux # default is 
-#        goarch: amd64 # default is 
+        goarch: amd64 # default is 
-#        build_flags: -tags 5BytesOffset # optional, default is 
+        release_tag: dev
-#        ldflags: -extldflags -static -X github.com/chrislusf/seaweedfs/weed/util.COMMIT=${{github.sha}}
+        overwrite: true 
-#        # Where to run `go build .`
+        build_flags: -tags 5BytesOffset # optional, default is 
-#        project_path: weed
+        ldflags: -extldflags -static -X github.com/chrislusf/seaweedfs/weed/util.COMMIT=${{github.sha}}
-#        binary_name: weed-large-disk
+        # Where to run `go build .`
+        project_path: weed
+        binary_name: weed-large-disk
+    - name: Go Release Binaries
+      uses: wangyoucao577/go-release-action@v1.10
+      with:
+        github_token: ${{ secrets.GITHUB_TOKEN }}
+        goos: linux # default is 
+        goarch: amd64 # default is 
+        release_tag: dev
+        overwrite: true 
+        ldflags: -extldflags -static -X github.com/chrislusf/seaweedfs/weed/util.COMMIT=${{github.sha}}
+        # Where to run `go build .`
+        project_path: weed
+        binary_name: weed

README.md

@@ -58,11 +58,12 @@ Your support will be really appreciated by me and other supporters!
 Table of Contents
 =================
 
+* [Quick Start](#quick-start)
 * [Introduction](#introduction)
 * [Features](#features)
     * [Additional Features](#additional-features)
     * [Filer Features](#filer-features)
-* [Example Usage](#example-usage)
+* [Example: Using Seaweed Object Store](#example-Using-Seaweed-Object-Store)
 * [Architecture](#architecture)
 * [Compared to Other File Systems](#compared-to-other-file-systems)
     * [Compared to HDFS](#compared-to-hdfs)
@@ -75,6 +76,13 @@ Table of Contents
 * [Benchmark](#Benchmark)
 * [License](#license)
 
+
+## Quick Start ##
+* Download the latest binary from https://github.com/chrislusf/seaweedfs/releases and unzip a single binary file `weed` or `weed.exe`
+* Run `weed server -dir=/some/data/dir -s3` to start one master, one volume server, one filer, and one S3 gateway.
+
+Also, to increase capacity, just add more volume servers by `weed volume -dir="/some/data/dir2" -mserver="<master_host>:9333" -port=8081` locally or a different machine. That is it!
+
 ## Introduction ##
 
 SeaweedFS is a simple and highly scalable distributed file system. There are two objectives:
@@ -141,10 +149,6 @@ On top of the object store, optional [Filer] can support directories and POSIX a
 
 [Back to TOC](#table-of-contents)
 
-## Quick Start ##
-* Download the latest binary from https://github.com/chrislusf/seaweedfs/releases and unzip a single binary file `weed` or `weed.exe`
-* Run `weed server -s3` to start one master, one volume server, one filer, and one S3 gateway.
-
 ## Example: Using Seaweed Object Store ##
 
 By default, the master node runs on port 9333, and the volume nodes run on port 8080.

docker/Makefile

@@ -12,6 +12,9 @@ build:
 dev: build
 	docker-compose -f local-dev-compose.yml -p seaweedfs up
 
+dev_registry: build
+	docker-compose -f local-registry-compose.yml -p seaweedfs up
+
 cluster: build
 	docker-compose -f local-cluster-compose.yml -p seaweedfs up
 

docker/local-registry-compose.yml

@@ -0,0 +1,95 @@
+version: '2'
+
+services:
+  master:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 9333:9333
+      - 19333:19333
+    command: "master -ip=master -volumeSizeLimitMB=1024"
+  volume:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8080:8080
+      - 18080:18080
+    command: "volume -mserver=master:9333 -port=8080 -ip=volume -max=0 -preStopSeconds=1"
+    volumes:
+      - type: bind
+        source: /Volumes/mobile_disk/data
+        target: /data
+    depends_on:
+      - master
+  filer:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8888:8888
+      - 18888:18888
+    command: 'filer -master="master:9333"'
+    depends_on:
+      - master
+      - volume
+  s3:
+    image: chrislusf/seaweedfs:local
+    ports:
+      - 8333:8333
+    command: '-v 9 s3 -filer="filer:8888"'
+    depends_on:
+      - master
+      - volume
+      - filer
+  minio:
+    image: minio/minio
+    ports:
+      - 9000:9000
+    command: 'minio server /data'
+    environment:
+      MINIO_ACCESS_KEY: "some_access_key1"
+      MINIO_SECRET_KEY: "some_secret_key1"
+    depends_on:
+      - master
+  registry1:
+    image: registry:2
+    environment:
+      REGISTRY_HTTP_ADDR: "0.0.0.0:5001" # seaweedfs s3
+      REGISTRY_LOG_LEVEL: "debug"
+      REGISTRY_STORAGE: "s3"
+      REGISTRY_STORAGE_S3_REGION: "us-east-1"
+      REGISTRY_STORAGE_S3_REGIONENDPOINT: "http://s3:8333"
+      REGISTRY_STORAGE_S3_BUCKET: "registry"
+      REGISTRY_STORAGE_S3_ACCESSKEY: "some_access_key1"
+      REGISTRY_STORAGE_S3_SECRETKEY: "some_secret_key1"
+      REGISTRY_STORAGE_S3_V4AUTH: "true"
+      REGISTRY_STORAGE_S3_SECURE: "false"
+      REGISTRY_STORAGE_S3_SKIPVERIFY: "true"
+      REGISTRY_STORAGE_S3_ROOTDIRECTORY: "/"
+      REGISTRY_STORAGE_DELETE_ENABLED: "true"
+      REGISTRY_STORAGE_REDIRECT_DISABLE: "true"
+      REGISTRY_VALIDATION_DISABLED: "true"
+    ports:
+      - 5001:5001
+    depends_on:
+      - s3
+      - minio
+  registry2:
+    image: registry:2
+    environment:
+      REGISTRY_HTTP_ADDR: "0.0.0.0:5002" # minio
+      REGISTRY_LOG_LEVEL: "debug"
+      REGISTRY_STORAGE: "s3"
+      REGISTRY_STORAGE_S3_REGION: "us-east-1"
+      REGISTRY_STORAGE_S3_REGIONENDPOINT: "http://minio:9000"
+      REGISTRY_STORAGE_S3_BUCKET: "registry"
+      REGISTRY_STORAGE_S3_ACCESSKEY: "some_access_key1"
+      REGISTRY_STORAGE_S3_SECRETKEY: "some_secret_key1"
+      REGISTRY_STORAGE_S3_V4AUTH: "true"
+      REGISTRY_STORAGE_S3_SECURE: "false"
+      REGISTRY_STORAGE_S3_SKIPVERIFY: "true"
+      REGISTRY_STORAGE_S3_ROOTDIRECTORY: "/"
+      REGISTRY_STORAGE_DELETE_ENABLED: "true"
+      REGISTRY_STORAGE_REDIRECT_DISABLE: "true"
+      REGISTRY_VALIDATION_DISABLED: "true"
+    ports:
+      - 5002:5002
+    depends_on:
+      - s3
+      - minio

go.mod

@@ -11,7 +11,7 @@ require (
 	github.com/aws/aws-sdk-go v1.33.5
 	github.com/buraksezer/consistent v0.0.0-20191006190839-693edf70fd72
 	github.com/cespare/xxhash v1.1.0
-	github.com/chrislusf/raft v1.0.2-0.20201002174524-b13c3bfdb011
+	github.com/chrislusf/raft v1.0.3
 	github.com/coreos/go-semver v0.3.0 // indirect
 	github.com/dgrijalva/jwt-go v3.2.0+incompatible
 	github.com/disintegration/imaging v1.6.2
@@ -24,6 +24,7 @@ require (
 	github.com/facebookgo/subset v0.0.0-20200203212716-c811ad88dec4 // indirect
 	github.com/fclairamb/ftpserverlib v0.8.0
 	github.com/frankban/quicktest v1.7.2 // indirect
+	github.com/go-errors/errors v1.1.1 // indirect
 	github.com/go-redis/redis v6.15.7+incompatible
 	github.com/go-sql-driver/mysql v1.5.0
 	github.com/gocql/gocql v0.0.0-20190829130954-e163eff7a8c6
@@ -45,6 +46,7 @@ require (
 	github.com/klauspost/reedsolomon v1.9.2
 	github.com/kurin/blazer v0.5.3
 	github.com/lib/pq v1.2.0
+	github.com/lunixbochs/vtclean v1.0.0 // indirect
 	github.com/magiconair/properties v1.8.1 // indirect
 	github.com/mattn/go-ieproxy v0.0.0-20190805055040-f9202b1cfdeb // indirect
 	github.com/mattn/go-runewidth v0.0.4 // indirect
@@ -67,6 +69,9 @@ require (
 	github.com/tidwall/gjson v1.3.2
 	github.com/tidwall/match v1.0.1
 	github.com/valyala/bytebufferpool v1.0.0
+	github.com/viant/assertly v0.5.4 // indirect
+	github.com/viant/ptrie v0.3.0
+	github.com/viant/toolbox v0.33.2 // indirect
 	github.com/willf/bitset v1.1.10 // indirect
 	github.com/willf/bloom v2.0.3+incompatible
 	github.com/wsxiaoys/terminal v0.0.0-20160513160801-0940f3fc43a0 // indirect

go.sum

@@ -84,8 +84,8 @@ github.com/cespare/xxhash v1.1.0 h1:a6HrQnmkObjyL+Gs60czilIUGqrzKutQD6XZog3p+ko=
 github.com/cespare/xxhash v1.1.0/go.mod h1:XrSqR1VqqWfGrhpAt58auRo0WTKS1nRRg3ghfAqPWnc=
 github.com/cespare/xxhash/v2 v2.1.1 h1:6MnRN8NT7+YBpUIWxHtefFZOKTAPgGjpQSxqLNn0+qY=
 github.com/cespare/xxhash/v2 v2.1.1/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
-github.com/chrislusf/raft v1.0.2-0.20201002174524-b13c3bfdb011 h1:vN1GvfLgDg8kIPCdhuVKAjlYpxG1B86jiKejB6MC/Q0=
+github.com/chrislusf/raft v1.0.3 h1:11YrnzJtVa5z7m9lhY2p8VcPHoUlC1UswyoAo+U1m1k=
-github.com/chrislusf/raft v1.0.2-0.20201002174524-b13c3bfdb011/go.mod h1:Ep5DP+mJSosjfKiix1uU7Lc2Df/SX4oGJEpZlXH5l68=
+github.com/chrislusf/raft v1.0.3/go.mod h1:Ep5DP+mJSosjfKiix1uU7Lc2Df/SX4oGJEpZlXH5l68=
 github.com/clbanning/x2j v0.0.0-20191024224557-825249438eec/go.mod h1:jMjuTZXRI4dUb/I5gc9Hdhagfvm9+RyrPryS/auMzxE=
 github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
 github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
@@ -159,6 +159,8 @@ github.com/fsnotify/fsnotify v1.4.7 h1:IXs+QLmnXW2CcXuY+8Mzv/fWEsPGWxqefPtCP5CnV
 github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
 github.com/ghodss/yaml v1.0.0 h1:wQHKEahhL6wmXdzwWG11gIVCkOv05bNOh+Rxn0yngAk=
 github.com/ghodss/yaml v1.0.0/go.mod h1:4dBDuWmgqj2HViK6kFavaiC9ZROes6MMH2rRYeMEF04=
+github.com/go-errors/errors v1.1.1 h1:ljK/pL5ltg3qoN+OtN6yCv9HWSfMwxSx90GJCZQxYNg=
+github.com/go-errors/errors v1.1.1/go.mod h1:psDX2osz5VnTOnFWbDeWwS7yejl+uV3FEWEp4lssFEs=
 github.com/go-ini/ini v1.25.4/go.mod h1:ByCAeIL28uOIIG0E3PJtZPDL8WnHpFKFOtgjp+3Ies8=
 github.com/go-kit/kit v0.8.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
 github.com/go-kit/kit v0.9.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
@@ -386,6 +388,8 @@ github.com/lib/pq v1.2.0 h1:LXpIM/LZ5xGFhOpXAQUIMM1HdyqzVYM13zNdjCEEcA0=
 github.com/lib/pq v1.2.0/go.mod h1:5WUZQaWbwv1U+lTReE5YruASi9Al49XbQIvNi/34Woo=
 github.com/lightstep/lightstep-tracer-common/golang/gogo v0.0.0-20190605223551-bc2310a04743/go.mod h1:qklhhLq1aX+mtWk9cPHPzaBjWImj5ULL6C7HFJtXQMM=
 github.com/lightstep/lightstep-tracer-go v0.18.1/go.mod h1:jlF1pusYV4pidLvZ+XD0UBX0ZE6WURAspgAczcDHrL4=
+github.com/lunixbochs/vtclean v1.0.0 h1:xu2sLAri4lGiovBDQKxl5mrXyESr3gUr5m5SM5+LVb8=
+github.com/lunixbochs/vtclean v1.0.0/go.mod h1:pHhQNgMf3btfWnGBVipUOjRYhoOsdGqdm/+2c2E2WMI=
 github.com/lyft/protoc-gen-validate v0.0.13/go.mod h1:XbGvPuh87YZc5TdIa2/I4pLk0QoUACkjt2znoq26NVQ=
 github.com/magiconair/properties v1.8.0 h1:LLgXmsheXeRoUOBOjtwPQCWIYqM/LU1ayDtDePerRcY=
 github.com/magiconair/properties v1.8.0/go.mod h1:PppfXfuXeibc/6YijjN8zIbojt8czPbwD3XqdrwzmxQ=
@@ -621,6 +625,12 @@ github.com/urfave/cli v1.20.0/go.mod h1:70zkFmudgCuE/ngEzBv17Jvp/497gISqfk5gWijb
 github.com/urfave/cli v1.22.1/go.mod h1:Gos4lmkARVdJ6EkW0WaNv/tZAAMe9V7XWyB60NtXRu0=
 github.com/valyala/bytebufferpool v1.0.0 h1:GqA5TC/0021Y/b9FG4Oi9Mr3q7XYx6KllzawFIhcdPw=
 github.com/valyala/bytebufferpool v1.0.0/go.mod h1:6bBcMArwyJ5K/AmCkWv1jt77kVWyCJ6HpOuEn7z0Csc=
+github.com/viant/assertly v0.5.4 h1:5Hh4U3pLZa6uhCFAGpYOxck/8l9TZczEzoHNfJAhHEQ=
+github.com/viant/assertly v0.5.4/go.mod h1:aGifi++jvCrUaklKEKT0BU95igDNaqkvz+49uaYMPRU=
+github.com/viant/ptrie v0.3.0 h1:SDaRd7Gqr1+ItCNz0GpTxRdK21nOfqjV6YtBm9jGlMY=
+github.com/viant/ptrie v0.3.0/go.mod h1:VguMnbGfz95Zw+V5VarYSqtqslDxJbOv++xLzxkMhec=
+github.com/viant/toolbox v0.33.2 h1:Av844IIeGz81gT672qZemyptGfbrcxqGymA5RFnIPjE=
+github.com/viant/toolbox v0.33.2/go.mod h1:OxMCG57V0PXuIP2HNQrtJf2CjqdmbrOx5EkMILuUhzM=
 github.com/willf/bitset v1.1.10 h1:NotGKqX0KwQ72NUzqrjZq5ipPNDQex9lo3WpaS8L2sc=
 github.com/willf/bitset v1.1.10/go.mod h1:RjeCKbqT1RxIR/KWY6phxZiaY1IyutSBfGjNPySAYV4=
 github.com/willf/bloom v2.0.3+incompatible h1:QDacWdqcAUI1MPOwIQZRy9kOR7yxfyEmxX8Wdm2/JPA=

k8s/seaweedfs/Chart.yaml

@@ -1,4 +1,4 @@
 apiVersion: v1
 description: SeaweedFS
 name: seaweedfs
-version: 2.08
+version: 2.11

k8s/seaweedfs/templates/filer-statefulset.yaml

@@ -101,7 +101,7 @@ spec:
               -port={{ .Values.filer.port }} \
               {{- if .Values.filer.metricsPort }}
               -metricsPort {{ .Values.filer.metricsPort }} \
-              {{- end }}}
+              {{- end }}
               {{- if .Values.filer.redirectOnRead }}
               -redirectOnRead \
               {{- end }}

k8s/seaweedfs/templates/s3-deployment.yaml

@@ -73,7 +73,7 @@ spec:
               -port={{ .Values.s3.port }} \
               {{- if .Values.s3.metricsPort }}
               -metricsPort {{ .Values.s3.metricsPort }} \
-              {{- end }}}
+              {{- end }}
               {{- if .Values.global.enableSecurity }}
               -cert.file=/usr/local/share/ca-certificates/client/tls.crt \
               -key.file=/usr/local/share/ca-certificates/client/tls.key \

k8s/seaweedfs/templates/s3-service.yaml

@@ -14,10 +14,10 @@ spec:
     protocol: TCP
 {{- if .Values.s3.metricsPort }}
   - name: "swfs-s3-metrics"
-    port: {{ .Values.filer.s3 }}
+    port: {{ .Values.s3.metricsPort }}
     targetPort: {{ .Values.s3.metricsPort }}
     protocol: TCP
-{{- end }}}
+{{- end }}
   selector:
     app: {{ template "seaweedfs.name" . }}
     component: s3

k8s/seaweedfs/templates/s3-servicemonitor.yaml

@@ -15,4 +15,4 @@ spec:
   selector:
     app: {{ template "seaweedfs.name" . }}
     component: s3
-{{- end }}}
+{{- end }}

k8s/seaweedfs/templates/seaweedfs-grafana-dashboard.yaml

@@ -0,0 +1,20 @@
+{{- if .Values.global.monitoring.enabled }}
+{{- $files := .Files.Glob "dashboards/*.json" }}
+{{- if $files }}
+apiVersion: v1
+kind: ConfigMapList
+items:
+{{- range $path, $fileContents := $files }}
+{{- $dashboardName := regexReplaceAll "(^.*/)(.*)\\.json$" $path "${2}" }}
+- apiVersion: v1
+  kind: ConfigMap
+  metadata:
+    name: {{ printf "%s" $dashboardName | lower | replace "_" "-" }}
+    namespace: {{ $.Release.Namespace }}
+    labels:
+     grafana_dashboard: "1"
+  data:
+    {{ $dashboardName }}.json: {{ $.Files.Get $path | toJson }}
+{{- end }}
+{{- end }}
+{{- end }}

k8s/seaweedfs/templates/volume-service.yaml

@@ -22,7 +22,7 @@ spec:
     port: {{ .Values.volume.metricsPort }}
     targetPort: {{ .Values.volume.metricsPort }}
     protocol: TCP
-{{- end }}}
+{{- end }}
 selector:
     app: {{ template "seaweedfs.name" . }}
     component: volume

k8s/seaweedfs/templates/volume-servicemonitor.yaml

@@ -15,4 +15,4 @@ spec:
   selector:
     app: {{ template "seaweedfs.name" . }}
     component: volume
-{{- end }}}
+{{- end }}

k8s/seaweedfs/templates/volume-statefulset.yaml

@@ -78,7 +78,7 @@ spec:
                 -port={{ .Values.volume.port }} \
                 {{- if .Values.volume.metricsPort }}
                 -metricsPort {{ .Values.volume.metricsPort }} \
-                {{- end }}}
+                {{- end }}
                 -dir={{ .Values.volume.dir }} \
                 -max={{ .Values.volume.maxVolumes }} \
                 {{- if .Values.volume.rack }}

k8s/seaweedfs/values.yaml

@@ -4,7 +4,7 @@ global:
   registry: ""
   repository: ""
   imageName: chrislusf/seaweedfs
-  imageTag: "2.08"
+  imageTag: "2.11"
   imagePullPolicy: IfNotPresent
   imagePullSecrets: imagepullsecret
   restartPolicy: Always
@@ -302,10 +302,6 @@ filer:
     WEED_FILER_OPTIONS_RECURSIVE_DELETE: "false"
     # directories under this folder will be automatically creating a separate bucket
     WEED_FILER_BUCKETS_FOLDER: "/buckets"
-    # directories under this folder will be store message queue data
-    WEED_FILER_QUEUES_FOLDER: "/queues"
-    # WEED_FILER_OPTIONS_BUCKETS_FSYNC a list of buckets names with all write requests fsync=true
-    WEED_FILER_OPTIONS_BUCKETS_FSYNC: []
 
 s3:
   enabled: true

other/java/client/pom.xml

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.5.2</version>
+    <version>1.5.6</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>
@@ -68,6 +68,11 @@
             <version>4.13.1</version>
             <scope>test</scope>
         </dependency>
+        <dependency>
+            <groupId>javax.annotation</groupId>
+            <artifactId>javax.annotation-api</artifactId>
+            <version>1.3.2</version>
+        </dependency>
     </dependencies>
 
     <distributionManagement>

other/java/client/pom.xml.deploy

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.5.2</version>
+    <version>1.5.6</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>

other/java/client/pom_debug.xml

@@ -5,7 +5,7 @@
 
     <groupId>com.github.chrislusf</groupId>
     <artifactId>seaweedfs-client</artifactId>
-    <version>1.5.2</version>
+    <version>1.5.6</version>
 
     <parent>
         <groupId>org.sonatype.oss</groupId>

other/java/client/src/main/java/seaweedfs/client/ChunkCache.java

@@ -15,7 +15,6 @@ public class ChunkCache {
         }
         this.cache = CacheBuilder.newBuilder()
                 .maximumSize(maxEntries)
-                .weakValues()
                 .expireAfterAccess(1, TimeUnit.HOURS)
                 .build();
     }

other/java/client/src/main/java/seaweedfs/client/FilerClient.java

@@ -333,4 +333,13 @@ public class FilerClient {
         return true;
     }
 
+    public Iterator<FilerProto.SubscribeMetadataResponse> watch(String prefix, String clientName, long sinceNs) {
+        return filerGrpcClient.getBlockingStub().subscribeMetadata(FilerProto.SubscribeMetadataRequest.newBuilder()
+                .setPathPrefix(prefix)
+                .setClientName(clientName)
+                .setSinceNs(sinceNs)
+                .build()
+        );
+    }
+
 }

other/java/client/src/main/java/seaweedfs/client/SeaweedRead.java

@@ -19,6 +19,7 @@ public class SeaweedRead {
     private static final Logger LOG = LoggerFactory.getLogger(SeaweedRead.class);
 
     static ChunkCache chunkCache = new ChunkCache(4);
+    static VolumeIdCache volumeIdCache = new VolumeIdCache(4 * 1024);
 
     // returns bytesRead
     public static long read(FilerGrpcClient filerGrpcClient, List<VisibleInterval> visibleIntervals,
@@ -27,16 +28,28 @@ public class SeaweedRead {
 
         List<ChunkView> chunkViews = viewFromVisibles(visibleIntervals, position, bufferLength);
 
+        Map<String, FilerProto.Locations> knownLocations = new HashMap<>();
+
         FilerProto.LookupVolumeRequest.Builder lookupRequest = FilerProto.LookupVolumeRequest.newBuilder();
         for (ChunkView chunkView : chunkViews) {
             String vid = parseVolumeId(chunkView.fileId);
+            FilerProto.Locations locations = volumeIdCache.getLocations(vid);
+            if (locations == null) {
                 lookupRequest.addVolumeIds(vid);
+            } else {
+                knownLocations.put(vid, locations);
+            }
         }
 
+        if (lookupRequest.getVolumeIdsCount() > 0) {
             FilerProto.LookupVolumeResponse lookupResponse = filerGrpcClient
                     .getBlockingStub().lookupVolume(lookupRequest.build());
-
             Map<String, FilerProto.Locations> vid2Locations = lookupResponse.getLocationsMapMap();
+            for (Map.Entry<String, FilerProto.Locations> entry : vid2Locations.entrySet()) {
+                volumeIdCache.setLocations(entry.getKey(), entry.getValue());
+                knownLocations.put(entry.getKey(), entry.getValue());
+            }
+        }
 
         //TODO parallel this
         long readCount = 0;
@@ -50,7 +63,7 @@ public class SeaweedRead {
                 startOffset += gap;
             }
 
-            FilerProto.Locations locations = vid2Locations.get(parseVolumeId(chunkView.fileId));
+            FilerProto.Locations locations = knownLocations.get(parseVolumeId(chunkView.fileId));
             if (locations == null || locations.getLocationsCount() == 0) {
                 LOG.error("failed to locate {}", chunkView.fileId);
                 // log here!

other/java/client/src/main/java/seaweedfs/client/VolumeIdCache.java

@@ -0,0 +1,36 @@
+package seaweedfs.client;
+
+import com.google.common.cache.Cache;
+import com.google.common.cache.CacheBuilder;
+
+import java.util.concurrent.TimeUnit;
+
+public class VolumeIdCache {
+
+    private Cache<String, FilerProto.Locations> cache = null;
+
+    public VolumeIdCache(int maxEntries) {
+        if (maxEntries == 0) {
+            return;
+        }
+        this.cache = CacheBuilder.newBuilder()
+                .maximumSize(maxEntries)
+                .expireAfterAccess(5, TimeUnit.MINUTES)
+                .build();
+    }
+
+    public FilerProto.Locations getLocations(String volumeId) {
+        if (this.cache == null) {
+            return null;
+        }
+        return this.cache.getIfPresent(volumeId);
+    }
+
+    public void setLocations(String volumeId, FilerProto.Locations locations) {
+        if (this.cache == null) {
+            return;
+        }
+        this.cache.put(volumeId, locations);
+    }
+
+}

other/java/client/src/main/proto/filer.proto

@@ -348,3 +348,23 @@ message KvPutRequest {
 message KvPutResponse {
     string error = 1;
 }
+
+// path-based configurations
+message FilerConf {
+    int32 version = 1;
+    message PathConf {
+        string location_prefix = 1;
+        string collection = 2;
+        string replication = 3;
+        string ttl = 4;
+        enum DiskType {
+            NONE = 0;
+            HDD = 1;
+            SSD = 2;
+        }
+        DiskType disk_type = 5;
+        bool fsync = 6;
+        uint32 volume_growth_count = 7;
+    }
+    repeated PathConf locations = 2;
+}

other/java/examples/pom.xml

@@ -0,0 +1,32 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project xmlns="http://maven.apache.org/POM/4.0.0"
+         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+    <modelVersion>4.0.0</modelVersion>
+
+    <groupId>org.example</groupId>
+    <artifactId>unzip</artifactId>
+    <version>1.0-SNAPSHOT</version>
+    <dependencies>
+        <dependency>
+            <groupId>com.github.chrislusf</groupId>
+            <artifactId>seaweedfs-client</artifactId>
+            <version>1.5.6</version>
+            <scope>compile</scope>
+        </dependency>
+        <dependency>
+            <groupId>com.github.chrislusf</groupId>
+            <artifactId>seaweedfs-hadoop2-client</artifactId>
+            <version>1.5.6</version>
+            <scope>compile</scope>
+        </dependency>
+        <dependency>
+            <groupId>org.apache.hadoop</groupId>
+            <artifactId>hadoop-common</artifactId>
+            <version>2.9.2</version>
+            <scope>compile</scope>
+        </dependency>
+    </dependencies>
+
+
+</project>

other/java/examples/src/main/java/com/seaweedfs/examples/UnzipFile.java

@@ -0,0 +1,54 @@
+package com.seaweedfs.examples;
+
+import seaweed.hdfs.SeaweedInputStream;
+import seaweedfs.client.FilerClient;
+import seaweedfs.client.FilerGrpcClient;
+
+import java.io.FileInputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipInputStream;
+
+public class UnzipFile {
+
+    public static void main(String[] args) throws IOException {
+
+        FilerGrpcClient filerGrpcClient = new FilerGrpcClient("localhost", 18888);
+        FilerClient filerClient = new FilerClient(filerGrpcClient);
+
+        long startTime = System.currentTimeMillis();
+        parseZip("/Users/chris/tmp/test.zip");
+
+        long startTime2 = System.currentTimeMillis();
+
+        long localProcessTime = startTime2 - startTime;
+
+        SeaweedInputStream seaweedInputStream = new SeaweedInputStream(
+                filerGrpcClient,
+                new org.apache.hadoop.fs.FileSystem.Statistics(""),
+                "/",
+                filerClient.lookupEntry("/", "test.zip")
+        );
+        parseZip(seaweedInputStream);
+
+        long swProcessTime = System.currentTimeMillis() - startTime2;
+
+        System.out.println("Local time: " + localProcessTime);
+        System.out.println("SeaweedFS time: " + swProcessTime);
+
+    }
+
+    public static void parseZip(String filename) throws IOException {
+        FileInputStream fileInputStream = new FileInputStream(filename);
+        parseZip(fileInputStream);
+    }
+
+    public static void parseZip(InputStream is) throws IOException {
+        ZipInputStream zin = new ZipInputStream(is);
+        ZipEntry ze;
+        while ((ze = zin.getNextEntry()) != null) {
+            System.out.println(ze.getName());
+        }
+    }
+}

other/java/examples/src/main/java/com/seaweedfs/examples/WatchFiles.java

@@ -0,0 +1,46 @@
+package com.seaweedfs.examples;
+
+import seaweedfs.client.FilerClient;
+import seaweedfs.client.FilerProto;
+
+import java.io.IOException;
+import java.util.Date;
+import java.util.Iterator;
+
+public class WatchFiles {
+
+    public static void main(String[] args) throws IOException {
+
+        FilerClient filerClient = new FilerClient("localhost", 18888);
+
+        long sinceNs = (System.currentTimeMillis() - 3600 * 1000) * 1000000L;
+
+        Iterator<FilerProto.SubscribeMetadataResponse> watch = filerClient.watch(
+                "/buckets",
+                "exampleClientName",
+                sinceNs
+        );
+
+        System.out.println("Connected to filer, subscribing from " + new Date());
+
+        while (watch.hasNext()) {
+            FilerProto.SubscribeMetadataResponse event = watch.next();
+            FilerProto.EventNotification notification = event.getEventNotification();
+            if (!event.getDirectory().equals(notification.getNewParentPath())) {
+                // move an entry to a new directory, possibly with a new name
+                if (notification.hasOldEntry() && notification.hasNewEntry()) {
+                    System.out.println("moved " + event.getDirectory() + "/" + notification.getOldEntry().getName() + " to " + notification.getNewParentPath() + "/" + notification.getNewEntry().getName());
+                } else {
+                    System.out.println("this should not happen.");
+                }
+            } else if (notification.hasNewEntry() && !notification.hasOldEntry()) {
+                System.out.println("created entry " + event.getDirectory() + "/" + notification.getNewEntry().getName());
+            } else if (!notification.hasNewEntry() && notification.hasOldEntry()) {
+                System.out.println("deleted entry " + event.getDirectory() + "/" + notification.getOldEntry().getName());
+            } else if (notification.hasNewEntry() && notification.hasOldEntry()) {
+                System.out.println("updated entry " + event.getDirectory() + "/" + notification.getNewEntry().getName());
+            }
+        }
+
+    }
+}

other/java/hdfs2/dependency-reduced-pom.xml

@@ -301,7 +301,7 @@
     </snapshotRepository>
   </distributionManagement>
   <properties>
-    <seaweedfs.client.version>1.5.2</seaweedfs.client.version>
+    <seaweedfs.client.version>1.5.6</seaweedfs.client.version>
     <hadoop.version>2.9.2</hadoop.version>
   </properties>
 </project>

other/java/hdfs2/pom.xml

@@ -5,7 +5,7 @@
     <modelVersion>4.0.0</modelVersion>
 
     <properties>
-        <seaweedfs.client.version>1.5.2</seaweedfs.client.version>
+        <seaweedfs.client.version>1.5.6</seaweedfs.client.version>
         <hadoop.version>2.9.2</hadoop.version>
     </properties>
 

other/java/hdfs2/src/main/java/seaweed/hdfs/SeaweedFileSystem.java

@@ -25,6 +25,7 @@ public class SeaweedFileSystem extends FileSystem {
     public static final String FS_SEAWEED_FILER_PORT = "fs.seaweed.filer.port";
     public static final int FS_SEAWEED_DEFAULT_PORT = 8888;
     public static final String FS_SEAWEED_BUFFER_SIZE = "fs.seaweed.buffer.size";
+    public static final String FS_SEAWEED_REPLICATION = "fs.seaweed.replication";
     public static final int FS_SEAWEED_DEFAULT_BUFFER_SIZE = 4 * 1024 * 1024;
 
     private static final Logger LOG = LoggerFactory.getLogger(SeaweedFileSystem.class);
@@ -65,6 +66,12 @@ public class SeaweedFileSystem extends FileSystem {
 
     }
 
+    @Override
+    public void close() throws IOException {
+        super.close();
+        this.seaweedFileSystemStore.close();
+    }
+
     @Override
     public FSDataInputStream open(Path path, int bufferSize) throws IOException {
 
@@ -91,7 +98,7 @@ public class SeaweedFileSystem extends FileSystem {
         path = qualify(path);
 
         try {
-            String replicaPlacement = String.format("%03d", replication - 1);
+            String replicaPlacement = this.getConf().get(FS_SEAWEED_REPLICATION, String.format("%03d", replication - 1));
             int seaweedBufferSize = this.getConf().getInt(FS_SEAWEED_BUFFER_SIZE, FS_SEAWEED_DEFAULT_BUFFER_SIZE);
             OutputStream outputStream = seaweedFileSystemStore.createFile(path, overwrite, permission, seaweedBufferSize, replicaPlacement);
             return new FSDataOutputStream(outputStream, statistics);

other/java/hdfs2/src/main/java/seaweed/hdfs/SeaweedFileSystemStore.java

@@ -36,6 +36,14 @@ public class SeaweedFileSystemStore {
         this.conf = conf;
     }
 
+    public void close() {
+        try {
+            this.filerGrpcClient.shutdown();
+        } catch (InterruptedException e) {
+            e.printStackTrace();
+        }
+    }
+
     public static String getParentDirectory(Path path) {
         return path.isRoot() ? "/" : path.getParent().toUri().getPath();
     }

other/java/hdfs3/dependency-reduced-pom.xml

@@ -309,7 +309,7 @@
     </snapshotRepository>
   </distributionManagement>
   <properties>
-    <seaweedfs.client.version>1.5.2</seaweedfs.client.version>
+    <seaweedfs.client.version>1.5.6</seaweedfs.client.version>
     <hadoop.version>3.1.1</hadoop.version>
   </properties>
 </project>

other/java/hdfs3/pom.xml

@@ -5,7 +5,7 @@
     <modelVersion>4.0.0</modelVersion>
 
     <properties>
-        <seaweedfs.client.version>1.5.2</seaweedfs.client.version>
+        <seaweedfs.client.version>1.5.6</seaweedfs.client.version>
         <hadoop.version>3.1.1</hadoop.version>
     </properties>
 

other/java/hdfs3/src/main/java/seaweed/hdfs/SeaweedFileSystem.java

@@ -25,6 +25,7 @@ public class SeaweedFileSystem extends FileSystem {
     public static final String FS_SEAWEED_FILER_PORT = "fs.seaweed.filer.port";
     public static final int FS_SEAWEED_DEFAULT_PORT = 8888;
     public static final String FS_SEAWEED_BUFFER_SIZE = "fs.seaweed.buffer.size";
+    public static final String FS_SEAWEED_REPLICATION = "fs.seaweed.replication";
     public static final int FS_SEAWEED_DEFAULT_BUFFER_SIZE = 4 * 1024 * 1024;
 
     private static final Logger LOG = LoggerFactory.getLogger(SeaweedFileSystem.class);
@@ -65,6 +66,12 @@ public class SeaweedFileSystem extends FileSystem {
 
     }
 
+    @Override
+    public void close() throws IOException {
+        super.close();
+        this.seaweedFileSystemStore.close();
+    }
+
     @Override
     public FSDataInputStream open(Path path, int bufferSize) throws IOException {
 
@@ -91,7 +98,7 @@ public class SeaweedFileSystem extends FileSystem {
         path = qualify(path);
 
         try {
-            String replicaPlacement = String.format("%03d", replication - 1);
+            String replicaPlacement = this.getConf().get(FS_SEAWEED_REPLICATION, String.format("%03d", replication - 1));
             int seaweedBufferSize = this.getConf().getInt(FS_SEAWEED_BUFFER_SIZE, FS_SEAWEED_DEFAULT_BUFFER_SIZE);
             OutputStream outputStream = seaweedFileSystemStore.createFile(path, overwrite, permission, seaweedBufferSize, replicaPlacement);
             return new FSDataOutputStream(outputStream, statistics);

other/java/hdfs3/src/main/java/seaweed/hdfs/SeaweedFileSystemStore.java

@@ -36,6 +36,14 @@ public class SeaweedFileSystemStore {
         this.conf = conf;
     }
 
+    public void close() {
+        try {
+            this.filerGrpcClient.shutdown();
+        } catch (InterruptedException e) {
+            e.printStackTrace();
+        }
+    }
+
     public static String getParentDirectory(Path path) {
         return path.isRoot() ? "/" : path.getParent().toUri().getPath();
     }

unmaintained/s3/benchmark/hsbench.sh

@@ -0,0 +1,3 @@
+#!/bin/bash
+
+hsbench -a accesstoken -s secret -z 4K -d 10 -t 10 -b 10 -u http://localhost:8333 -m "cxipgdx" -bp "hsbench-"

weed/command/benchmark.go

@@ -125,7 +125,7 @@ func runBenchmark(cmd *Command, args []string) bool {
 		defer pprof.StopCPUProfile()
 	}
 
-	b.masterClient = wdclient.NewMasterClient(b.grpcDialOption, "client", "", 0, strings.Split(*b.masters, ","))
+	b.masterClient = wdclient.NewMasterClient(b.grpcDialOption, "client", "", 0, "", strings.Split(*b.masters, ","))
 	go b.masterClient.KeepConnectedToMaster()
 	b.masterClient.WaitUntilConnected()
 

weed/command/filer.go

@@ -59,7 +59,7 @@ func init() {
 	f.disableDirListing = cmdFiler.Flag.Bool("disableDirListing", false, "turn off directory listing")
 	f.maxMB = cmdFiler.Flag.Int("maxMB", 32, "split files larger than the limit")
 	f.dirListingLimit = cmdFiler.Flag.Int("dirListLimit", 100000, "limit sub dir listing size")
-	f.dataCenter = cmdFiler.Flag.String("dataCenter", "", "prefer to write to volumes in this data center")
+	f.dataCenter = cmdFiler.Flag.String("dataCenter", "", "prefer to read and write to volumes in this data center")
 	f.rack = cmdFiler.Flag.String("rack", "", "prefer to write to volumes in this rack")
 	f.disableHttp = cmdFiler.Flag.Bool("disableHttp", false, "disable http request, only gRpc operations are allowed")
 	f.cipher = cmdFiler.Flag.Bool("encryptVolumeData", false, "encrypt data on volume servers")

weed/command/filer_copy.go

@@ -122,7 +122,7 @@ func runCopy(cmd *Command, args []string) bool {
 			expectedBucket := restPath[:strings.Index(restPath, "/")]
 			if *copy.collection == "" {
 				*copy.collection = expectedBucket
-			} else {
+			} else if *copy.collection != expectedBucket {
 				fmt.Printf("destination %s uses collection \"%s\": unexpected collection \"%v\"\n", urlPath, expectedBucket, *copy.collection)
 				return true
 			}

weed/command/scaffold.go

@@ -76,10 +76,6 @@ const (
 recursive_delete = false
 # directories under this folder will be automatically creating a separate bucket
 buckets_folder = "/buckets"
-buckets_fsync = [          # a list of buckets with all write requests fsync=true
-	"important_bucket",
-	"should_always_fsync",
-]
 
 ####################################################
 # The following are filer store options

weed/filer/filer.go

@@ -41,14 +41,16 @@ type Filer struct {
 	metaLogReplication  string
 	MetaAggregator      *MetaAggregator
 	Signature           int32
+	FilerConf           *FilerConf
 }
 
 func NewFiler(masters []string, grpcDialOption grpc.DialOption,
-	filerHost string, filerGrpcPort uint32, collection string, replication string, notifyFn func()) *Filer {
+	filerHost string, filerGrpcPort uint32, collection string, replication string, dataCenter string, notifyFn func()) *Filer {
 	f := &Filer{
-		MasterClient:        wdclient.NewMasterClient(grpcDialOption, "filer", filerHost, filerGrpcPort, masters),
+		MasterClient:        wdclient.NewMasterClient(grpcDialOption, "filer", filerHost, filerGrpcPort, dataCenter, masters),
 		fileIdDeletionQueue: util.NewUnboundedQueue(),
 		GrpcDialOption:      grpcDialOption,
+		FilerConf:           NewFilerConf(),
 	}
 	f.LocalMetaLogBuffer = log_buffer.NewLogBuffer(LogFlushInterval, f.logFlushFunc, notifyFn)
 	f.metaLogCollection = collection

weed/filer/filer_conf.go

@@ -0,0 +1,148 @@
+package filer
+
+import (
+	"bytes"
+	"context"
+	"io"
+
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/util"
+	"github.com/golang/protobuf/jsonpb"
+	"github.com/golang/protobuf/proto"
+	"github.com/viant/ptrie"
+)
+
+const (
+	DirectoryEtc  = "/etc"
+	FilerConfName = "filer.conf"
+)
+
+type FilerConf struct {
+	rules ptrie.Trie
+}
+
+func NewFilerConf() (fc *FilerConf) {
+	fc = &FilerConf{
+		rules: ptrie.New(),
+	}
+	return fc
+}
+
+func (fc *FilerConf) loadFromFiler(filer *Filer) (err error) {
+	filerConfPath := util.NewFullPath(DirectoryEtc, FilerConfName)
+	entry, err := filer.FindEntry(context.Background(), filerConfPath)
+	if err != nil {
+		if err == filer_pb.ErrNotFound {
+			return nil
+		}
+		glog.Errorf("read filer conf entry %s: %v", filerConfPath, err)
+		return
+	}
+
+	return fc.loadFromChunks(filer, entry.Chunks)
+}
+
+func (fc *FilerConf) loadFromChunks(filer *Filer, chunks []*filer_pb.FileChunk) (err error) {
+	data, err := filer.readEntry(chunks)
+	if err != nil {
+		glog.Errorf("read filer conf content: %v", err)
+		return
+	}
+
+	return fc.LoadFromBytes(data)
+}
+
+func (fc *FilerConf) LoadFromBytes(data []byte) (err error) {
+	conf := &filer_pb.FilerConf{}
+
+	if err := jsonpb.Unmarshal(bytes.NewReader(data), conf); err != nil {
+
+		err = proto.UnmarshalText(string(data), conf)
+		if err != nil {
+			glog.Errorf("unable to parse filer conf: %v", err)
+			// this is not recoverable
+			return nil
+		}
+
+		return nil
+	}
+
+	return fc.doLoadConf(conf)
+}
+
+func (fc *FilerConf) doLoadConf(conf *filer_pb.FilerConf) (err error) {
+	for _, location := range conf.Locations {
+		err = fc.AddLocationConf(location)
+		if err != nil {
+			// this is not recoverable
+			return nil
+		}
+	}
+	return nil
+}
+
+func (fc *FilerConf) AddLocationConf(locConf *filer_pb.FilerConf_PathConf) (err error) {
+	err = fc.rules.Put([]byte(locConf.LocationPrefix), locConf)
+	if err != nil {
+		glog.Errorf("put location prefix: %v", err)
+	}
+	return
+}
+
+func (fc *FilerConf) DeleteLocationConf(locationPrefix string) {
+	rules := ptrie.New()
+	fc.rules.Walk(func(key []byte, value interface{}) bool {
+		if string(key) == locationPrefix {
+			return true
+		}
+		rules.Put(key, value)
+		return true
+	})
+	fc.rules = rules
+	return
+}
+
+func (fc *FilerConf) MatchStorageRule(path string) (pathConf *filer_pb.FilerConf_PathConf) {
+	pathConf = &filer_pb.FilerConf_PathConf{}
+	fc.rules.MatchPrefix([]byte(path), func(key []byte, value interface{}) bool {
+		t := value.(*filer_pb.FilerConf_PathConf)
+		mergePathConf(pathConf, t)
+		return true
+	})
+	return pathConf
+}
+
+// merge if values in b is not empty, merge them into a
+func mergePathConf(a, b *filer_pb.FilerConf_PathConf) {
+	a.Collection = util.Nvl(b.Collection, a.Collection)
+	a.Replication = util.Nvl(b.Replication, a.Replication)
+	a.Ttl = util.Nvl(b.Ttl, a.Ttl)
+	if b.DiskType != filer_pb.FilerConf_PathConf_NONE {
+		a.DiskType = b.DiskType
+	}
+	a.Fsync = b.Fsync || a.Fsync
+	if b.VolumeGrowthCount > 0 {
+		a.VolumeGrowthCount = b.VolumeGrowthCount
+	}
+}
+
+func (fc *FilerConf) ToProto() *filer_pb.FilerConf {
+	m := &filer_pb.FilerConf{}
+	fc.rules.Walk(func(key []byte, value interface{}) bool {
+		pathConf := value.(*filer_pb.FilerConf_PathConf)
+		m.Locations = append(m.Locations, pathConf)
+		return true
+	})
+	return m
+}
+
+func (fc *FilerConf) ToText(writer io.Writer) error {
+
+	m := jsonpb.Marshaler{
+		EmitDefaults: false,
+		Indent:       "  ",
+	}
+
+	return m.Marshal(writer, fc.ToProto())
+}

weed/filer/filer_conf_test.go

@@ -0,0 +1,34 @@
+package filer
+
+import (
+	"testing"
+
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestFilerConf(t *testing.T) {
+
+	fc := NewFilerConf()
+
+	conf := &filer_pb.FilerConf{Locations: []*filer_pb.FilerConf_PathConf{
+		{
+			LocationPrefix: "/buckets/abc",
+			Collection:     "abc",
+		},
+		{
+			LocationPrefix: "/buckets/abcd",
+			Collection:     "abcd",
+		},
+		{
+			LocationPrefix: "/buckets/",
+			Replication: "001",
+		},
+	}}
+	fc.doLoadConf(conf)
+
+	assert.Equal(t, "abc", fc.MatchStorageRule("/buckets/abc/jasdf").Collection)
+	assert.Equal(t, "abcd", fc.MatchStorageRule("/buckets/abcd/jasdf").Collection)
+	assert.Equal(t, "001", fc.MatchStorageRule("/buckets/abc/jasdf").Replication)
+
+}

weed/filer/filer_delete_entry.go

@@ -47,7 +47,7 @@ func (f *Filer) DeleteEntryMetaAndData(ctx context.Context, p util.FullPath, isR
 	}
 
 	if shouldDeleteChunks && !isCollection {
-		go f.DeleteChunks(chunks)
+		f.DirectDeleteChunks(chunks)
 	}
 	// A case not handled:
 	// what if the chunk is in a different collection?

weed/filer/filer_deletion.go

@@ -68,6 +68,50 @@ func (f *Filer) loopProcessingDeletion() {
 	}
 }
 
+func (f *Filer) doDeleteFileIds(fileIds []string) {
+
+	lookupFunc := LookupByMasterClientFn(f.MasterClient)
+	DeletionBatchSize := 100000 // roughly 20 bytes cost per file id.
+
+	for len(fileIds) > 0 {
+		var toDeleteFileIds []string
+		if len(fileIds) > DeletionBatchSize {
+			toDeleteFileIds = fileIds[:DeletionBatchSize]
+			fileIds = fileIds[DeletionBatchSize:]
+		} else {
+			toDeleteFileIds = fileIds
+			fileIds = fileIds[:0]
+		}
+		deletionCount := len(toDeleteFileIds)
+		_, err := operation.DeleteFilesWithLookupVolumeId(f.GrpcDialOption, toDeleteFileIds, lookupFunc)
+		if err != nil {
+			if !strings.Contains(err.Error(), "already deleted") {
+				glog.V(0).Infof("deleting fileIds len=%d error: %v", deletionCount, err)
+			}
+		}
+	}
+}
+
+func (f *Filer) DirectDeleteChunks(chunks []*filer_pb.FileChunk) {
+	var fildIdsToDelete []string
+	for _, chunk := range chunks {
+		if !chunk.IsChunkManifest {
+			fildIdsToDelete = append(fildIdsToDelete, chunk.GetFileIdString())
+			continue
+		}
+		dataChunks, manifestResolveErr := ResolveOneChunkManifest(f.MasterClient.LookupFileId, chunk)
+		if manifestResolveErr != nil {
+			glog.V(0).Infof("failed to resolve manifest %s: %v", chunk.FileId, manifestResolveErr)
+		}
+		for _, dChunk := range dataChunks {
+			fildIdsToDelete = append(fildIdsToDelete, dChunk.GetFileIdString())
+		}
+		fildIdsToDelete = append(fildIdsToDelete, chunk.GetFileIdString())
+	}
+
+	f.doDeleteFileIds(fildIdsToDelete)
+}
+
 func (f *Filer) DeleteChunks(chunks []*filer_pb.FileChunk) {
 	for _, chunk := range chunks {
 		if !chunk.IsChunkManifest {

weed/filer/filer_notify_append.go

@@ -13,7 +13,7 @@ import (
 
 func (f *Filer) appendToFile(targetFile string, data []byte) error {
 
-	assignResult, uploadResult, err2 := f.assignAndUpload(data)
+	assignResult, uploadResult, err2 := f.assignAndUpload(targetFile, data)
 	if err2 != nil {
 		return err2
 	}
@@ -46,14 +46,16 @@ func (f *Filer) appendToFile(targetFile string, data []byte) error {
 	return err
 }
 
-func (f *Filer) assignAndUpload(data []byte) (*operation.AssignResult, *operation.UploadResult, error) {
+func (f *Filer) assignAndUpload(targetFile string, data []byte) (*operation.AssignResult, *operation.UploadResult, error) {
 	// assign a volume location
+	rule := f.FilerConf.MatchStorageRule(targetFile)
 	assignRequest := &operation.VolumeAssignRequest{
 		Count:               1,
-		Collection:          f.metaLogCollection,
+		Collection:          util.Nvl(f.metaLogCollection, rule.Collection),
-		Replication:         f.metaLogReplication,
+		Replication:         util.Nvl(f.metaLogReplication, rule.Replication),
-		WritableVolumeCount: 1,
+		WritableVolumeCount: rule.VolumeGrowthCount,
 	}
+
 	assignResult, err := operation.Assign(f.GetMaster(), f.GrpcDialOption, assignRequest)
 	if err != nil {
 		return nil, nil, fmt.Errorf("AssignVolume: %v", err)

weed/filer/filer_on_meta_event.go

@@ -0,0 +1,61 @@
+package filer
+
+import (
+	"bytes"
+	"math"
+
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+// onMetadataChangeEvent is triggered after filer processed change events from local or remote filers
+func (f *Filer) onMetadataChangeEvent(event *filer_pb.SubscribeMetadataResponse) {
+	if DirectoryEtc != event.Directory {
+		if DirectoryEtc != event.EventNotification.NewParentPath {
+			return
+		}
+	}
+
+	entry := event.EventNotification.NewEntry
+	if entry == nil {
+		return
+	}
+
+	glog.V(0).Infof("procesing %v", event)
+	if entry.Name == FilerConfName {
+		f.reloadFilerConfiguration(entry)
+	}
+
+}
+
+func (f *Filer) readEntry(chunks []*filer_pb.FileChunk) ([]byte, error) {
+	var buf bytes.Buffer
+	err := StreamContent(f.MasterClient, &buf, chunks, 0, math.MaxInt64)
+	if err != nil {
+		return nil, err
+	}
+	return buf.Bytes(), nil
+}
+
+func (f *Filer) reloadFilerConfiguration(entry *filer_pb.Entry) {
+	fc := NewFilerConf()
+	err := fc.loadFromChunks(f, entry.Chunks)
+	if err != nil {
+		glog.Errorf("read filer conf chunks: %v", err)
+		return
+	}
+	f.FilerConf = fc
+}
+
+func (f *Filer) LoadFilerConf() {
+	fc := NewFilerConf()
+	err := util.Retry("loadFilerConf", func() error {
+		return fc.loadFromFiler(f)
+	})
+	if err != nil {
+		glog.Errorf("read filer conf: %v", err)
+		return
+	}
+	f.FilerConf = fc
+}

weed/filer/leveldb/leveldb_store_test.go

@@ -11,7 +11,7 @@ import (
 )
 
 func TestCreateAndFind(t *testing.T) {
-	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", nil)
+	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", "", nil)
 	dir, _ := ioutil.TempDir("", "seaweedfs_filer_test")
 	defer os.RemoveAll(dir)
 	store := &LevelDBStore{}
@@ -65,7 +65,7 @@ func TestCreateAndFind(t *testing.T) {
 }
 
 func TestEmptyRoot(t *testing.T) {
-	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", nil)
+	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", "", nil)
 	dir, _ := ioutil.TempDir("", "seaweedfs_filer_test2")
 	defer os.RemoveAll(dir)
 	store := &LevelDBStore{}

weed/filer/leveldb2/leveldb2_store_test.go

@@ -11,7 +11,7 @@ import (
 )
 
 func TestCreateAndFind(t *testing.T) {
-	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", nil)
+	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", "", nil)
 	dir, _ := ioutil.TempDir("", "seaweedfs_filer_test")
 	defer os.RemoveAll(dir)
 	store := &LevelDB2Store{}
@@ -65,7 +65,7 @@ func TestCreateAndFind(t *testing.T) {
 }
 
 func TestEmptyRoot(t *testing.T) {
-	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", nil)
+	testFiler := filer.NewFiler(nil, nil, "", 0, "", "", "", nil)
 	dir, _ := ioutil.TempDir("", "seaweedfs_filer_test2")
 	defer os.RemoveAll(dir)
 	store := &LevelDB2Store{}

weed/filer/meta_aggregator.go

@@ -141,6 +141,9 @@ func (ma *MetaAggregator) subscribeToOneFiler(f *Filer, self string, peer string
 					return fmt.Errorf("process %v: %v", resp, err)
 				}
 				lastTsNs = resp.TsNs
+
+				f.onMetadataChangeEvent(resp)
+
 			}
 		})
 		if err != nil {

weed/filesys/dirty_page.go

@@ -15,10 +15,10 @@ type ContinuousDirtyPages struct {
 	intervals              *ContinuousIntervals
 	f                      *File
 	writeWaitGroup         sync.WaitGroup
+	chunkAddLock           sync.Mutex
 	chunkSaveErrChan       chan error
 	chunkSaveErrChanClosed bool
 	lastErr                error
-	lock                   sync.Mutex
 	collection             string
 	replication            string
 }
@@ -117,6 +117,8 @@ func (pages *ContinuousDirtyPages) saveToStorage(reader io.Reader, offset int64,
 		}
 		chunk.Mtime = mtime
 		pages.collection, pages.replication = collection, replication
+		pages.chunkAddLock.Lock()
+		defer pages.chunkAddLock.Unlock()
 		pages.f.addChunks([]*filer_pb.FileChunk{chunk})
 		glog.V(3).Infof("%s saveToStorage [%d,%d)", pages.f.fullpath(), offset, offset+size)
 	}

weed/filesys/file.go

@@ -144,7 +144,7 @@ func (file *File) Setattr(ctx context.Context, req *fuse.SetattrRequest, resp *f
 				}
 			}
 			file.entry.Chunks = chunks
-			file.entryViewCache = nil
+			file.entryViewCache, _ = filer.NonOverlappingVisibleIntervals(filer.LookupFn(file.wfs), chunks)
 			file.reader = nil
 			file.wfs.deleteFileChunks(truncatedChunks)
 		}
@@ -282,15 +282,37 @@ func (file *File) maybeLoadEntry(ctx context.Context) (entry *filer_pb.Entry, er
 	return entry, nil
 }
 
+func lessThan(a, b *filer_pb.FileChunk) bool {
+	if a.Mtime == b.Mtime {
+		return a.Fid.FileKey < b.Fid.FileKey
+	}
+	return a.Mtime < b.Mtime
+}
+
 func (file *File) addChunks(chunks []*filer_pb.FileChunk) {
 
-	sort.Slice(chunks, func(i, j int) bool {
+	// find the earliest incoming chunk
-		if chunks[i].Mtime == chunks[j].Mtime {
+	newChunks := chunks
-			return chunks[i].Fid.FileKey < chunks[j].Fid.FileKey
+	earliestChunk := newChunks[0]
+	for i := 1; i < len(newChunks); i++ {
+		if lessThan(earliestChunk, newChunks[i]) {
+			earliestChunk = newChunks[i]
 		}
-		return chunks[i].Mtime < chunks[j].Mtime
+	}
+
+	// pick out-of-order chunks from existing chunks
+	for _, chunk := range file.entry.Chunks {
+		if lessThan(earliestChunk, chunk) {
+			chunks = append(chunks, chunk)
+		}
+	}
+
+	// sort incoming chunks
+	sort.Slice(chunks, func(i, j int) bool {
+		return lessThan(chunks[i], chunks[j])
 	})
 
+	// add to entry view cache
 	for _, chunk := range chunks {
 		file.entryViewCache = filer.MergeIntoVisibles(file.entryViewCache, chunk)
 	}
@@ -299,7 +321,7 @@ func (file *File) addChunks(chunks []*filer_pb.FileChunk) {
 
 	glog.V(4).Infof("%s existing %d chunks adds %d more", file.fullpath(), len(file.entry.Chunks), len(chunks))
 
-	file.entry.Chunks = append(file.entry.Chunks, chunks...)
+	file.entry.Chunks = append(file.entry.Chunks, newChunks...)
 }
 
 func (file *File) setEntry(entry *filer_pb.Entry) {

weed/filesys/filehandle.go

@@ -183,11 +183,10 @@ func (fh *FileHandle) Release(ctx context.Context, req *fuse.ReleaseRequest) err
 	}
 
 	if fh.f.isOpen == 0 {
+
 		if err := fh.doFlush(ctx, req.Header); err != nil {
 			glog.Errorf("Release doFlush %s: %v", fh.f.Name, err)
 		}
-		fh.f.wfs.ReleaseHandle(fh.f.fullpath(), fuse.HandleID(fh.handle))
-	}
 
 		// stop the goroutine
 		if !fh.dirtyPages.chunkSaveErrChanClosed {
@@ -195,6 +194,9 @@ func (fh *FileHandle) Release(ctx context.Context, req *fuse.ReleaseRequest) err
 			close(fh.dirtyPages.chunkSaveErrChan)
 		}
 
+		fh.f.wfs.ReleaseHandle(fh.f.fullpath(), fuse.HandleID(fh.handle))
+	}
+
 	return nil
 }
 
@@ -262,7 +264,6 @@ func (fh *FileHandle) doFlush(ctx context.Context, header fuse.Header) error {
 			glog.V(0).Infof("MaybeManifestize: %v", manifestErr)
 		}
 		fh.f.entry.Chunks = append(chunks, manifestChunks...)
-		fh.f.entryViewCache = nil
 
 		fh.f.wfs.mapPbIdFromLocalToFiler(request.Entry)
 		defer fh.f.wfs.mapPbIdFromFilerToLocal(request.Entry)

weed/operation/assign_file_id.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"strings"
 
+	"github.com/chrislusf/seaweedfs/weed/storage/needle"
 	"google.golang.org/grpc"
 
 	"github.com/chrislusf/seaweedfs/weed/pb/master_pb"
@@ -49,14 +50,14 @@ func Assign(server string, grpcDialOption grpc.DialOption, primaryRequest *Volum
 		lastError = WithMasterServerClient(server, grpcDialOption, func(masterClient master_pb.SeaweedClient) error {
 
 			req := &master_pb.AssignRequest{
-				Count:               primaryRequest.Count,
+				Count:               request.Count,
-				Replication:         primaryRequest.Replication,
+				Replication:         request.Replication,
-				Collection:          primaryRequest.Collection,
+				Collection:          request.Collection,
-				Ttl:                 primaryRequest.Ttl,
+				Ttl:                 request.Ttl,
-				DataCenter:          primaryRequest.DataCenter,
+				DataCenter:          request.DataCenter,
-				Rack:                primaryRequest.Rack,
+				Rack:                request.Rack,
-				DataNode:            primaryRequest.DataNode,
+				DataNode:            request.DataNode,
-				WritableVolumeCount: primaryRequest.WritableVolumeCount,
+				WritableVolumeCount: request.WritableVolumeCount,
 			}
 			resp, grpcErr := masterClient.Assign(context.Background(), req)
 			if grpcErr != nil {
@@ -101,3 +102,41 @@ func LookupJwt(master string, fileId string) security.EncodedJwt {
 
 	return security.EncodedJwt(tokenStr)
 }
+
+type StorageOption struct {
+	Replication       string
+	Collection        string
+	DataCenter        string
+	Rack              string
+	TtlSeconds        int32
+	Fsync             bool
+	VolumeGrowthCount uint32
+}
+
+func (so *StorageOption) TtlString() string {
+	return needle.SecondsToTTL(so.TtlSeconds)
+}
+
+func (so *StorageOption) ToAssignRequests(count int) (ar *VolumeAssignRequest, altRequest *VolumeAssignRequest) {
+	ar = &VolumeAssignRequest{
+		Count:               uint64(count),
+		Replication:         so.Replication,
+		Collection:          so.Collection,
+		Ttl:                 so.TtlString(),
+		DataCenter:          so.DataCenter,
+		Rack:                so.Rack,
+		WritableVolumeCount: so.VolumeGrowthCount,
+	}
+	if so.DataCenter != "" || so.Rack != "" {
+		altRequest = &VolumeAssignRequest{
+			Count:               uint64(count),
+			Replication:         so.Replication,
+			Collection:          so.Collection,
+			Ttl:                 so.TtlString(),
+			DataCenter:          "",
+			Rack:                "",
+			WritableVolumeCount: so.VolumeGrowthCount,
+		}
+	}
+	return
+}

weed/operation/upload_content.go

@@ -81,14 +81,11 @@ func doUpload(uploadUrl string, filename string, cipher bool, reader io.Reader,
 	if ok {
 		data = bytesReader.Bytes
 	} else {
-		buf := bytebufferpool.Get()
+		data, err = ioutil.ReadAll(reader)
-		_, err = buf.ReadFrom(reader)
-		defer bytebufferpool.Put(buf)
 		if err != nil {
 			err = fmt.Errorf("read input: %v", err)
 			return
 		}
-		data = buf.Bytes()
 	}
 	uploadResult, uploadErr := retriedUploadData(uploadUrl, filename, cipher, data, isInputCompressed, mtype, pairMap, jwt)
 	return uploadResult, uploadErr, data
@@ -172,7 +169,7 @@ func doUploadData(uploadUrl string, filename string, cipher bool, data []byte, i
 		uploadResult, err = upload_content(uploadUrl, func(w io.Writer) (err error) {
 			_, err = w.Write(data)
 			return
-		}, filename, contentIsGzipped, 0, mtype, pairMap, jwt)
+		}, filename, contentIsGzipped, len(data), mtype, pairMap, jwt)
 	}
 
 	if uploadResult == nil {
@@ -193,6 +190,7 @@ func upload_content(uploadUrl string, fillBufferFunction func(w io.Writer) error
 	body_writer := multipart.NewWriter(buf)
 	h := make(textproto.MIMEHeader)
 	h.Set("Content-Disposition", fmt.Sprintf(`form-data; name="file"; filename="%s"`, fileNameEscaper.Replace(filename)))
+	h.Set("Idempotency-Key", uploadUrl)
 	if mtype == "" {
 		mtype = mime.TypeByExtension(strings.ToLower(filepath.Ext(filename)))
 	}

weed/pb/filer.proto

@@ -348,3 +348,23 @@ message KvPutRequest {
 message KvPutResponse {
     string error = 1;
 }
+
+// path-based configurations
+message FilerConf {
+    int32 version = 1;
+    message PathConf {
+        string location_prefix = 1;
+        string collection = 2;
+        string replication = 3;
+        string ttl = 4;
+        enum DiskType {
+            NONE = 0;
+            HDD = 1;
+            SSD = 2;
+        }
+        DiskType disk_type = 5;
+        bool fsync = 6;
+        uint32 volume_growth_count = 7;
+    }
+    repeated PathConf locations = 2;
+}

weed/pb/filer_pb/filer.pb.go

@@ -29,6 +29,55 @@ const (
 // of the legacy proto package is being used.
 const _ = proto.ProtoPackageIsVersion4
 
+type FilerConf_PathConf_DiskType int32
+
+const (
+	FilerConf_PathConf_NONE FilerConf_PathConf_DiskType = 0
+	FilerConf_PathConf_HDD  FilerConf_PathConf_DiskType = 1
+	FilerConf_PathConf_SSD  FilerConf_PathConf_DiskType = 2
+)
+
+// Enum value maps for FilerConf_PathConf_DiskType.
+var (
+	FilerConf_PathConf_DiskType_name = map[int32]string{
+		0: "NONE",
+		1: "HDD",
+		2: "SSD",
+	}
+	FilerConf_PathConf_DiskType_value = map[string]int32{
+		"NONE": 0,
+		"HDD":  1,
+		"SSD":  2,
+	}
+)
+
+func (x FilerConf_PathConf_DiskType) Enum() *FilerConf_PathConf_DiskType {
+	p := new(FilerConf_PathConf_DiskType)
+	*p = x
+	return p
+}
+
+func (x FilerConf_PathConf_DiskType) String() string {
+	return protoimpl.X.EnumStringOf(x.Descriptor(), protoreflect.EnumNumber(x))
+}
+
+func (FilerConf_PathConf_DiskType) Descriptor() protoreflect.EnumDescriptor {
+	return file_filer_proto_enumTypes[0].Descriptor()
+}
+
+func (FilerConf_PathConf_DiskType) Type() protoreflect.EnumType {
+	return &file_filer_proto_enumTypes[0]
+}
+
+func (x FilerConf_PathConf_DiskType) Number() protoreflect.EnumNumber {
+	return protoreflect.EnumNumber(x)
+}
+
+// Deprecated: Use FilerConf_PathConf_DiskType.Descriptor instead.
+func (FilerConf_PathConf_DiskType) EnumDescriptor() ([]byte, []int) {
+	return file_filer_proto_rawDescGZIP(), []int{47, 0, 0}
+}
+
 type LookupDirectoryEntryRequest struct {
 	state         protoimpl.MessageState
 	sizeCache     protoimpl.SizeCache
@@ -3002,6 +3051,62 @@ func (x *KvPutResponse) GetError() string {
 	return ""
 }
 
+// path-based configurations
+type FilerConf struct {
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	Version   int32                 `protobuf:"varint,1,opt,name=version,proto3" json:"version,omitempty"`
+	Locations []*FilerConf_PathConf `protobuf:"bytes,2,rep,name=locations,proto3" json:"locations,omitempty"`
+}
+
+func (x *FilerConf) Reset() {
+	*x = FilerConf{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_filer_proto_msgTypes[47]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *FilerConf) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*FilerConf) ProtoMessage() {}
+
+func (x *FilerConf) ProtoReflect() protoreflect.Message {
+	mi := &file_filer_proto_msgTypes[47]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use FilerConf.ProtoReflect.Descriptor instead.
+func (*FilerConf) Descriptor() ([]byte, []int) {
+	return file_filer_proto_rawDescGZIP(), []int{47}
+}
+
+func (x *FilerConf) GetVersion() int32 {
+	if x != nil {
+		return x.Version
+	}
+	return 0
+}
+
+func (x *FilerConf) GetLocations() []*FilerConf_PathConf {
+	if x != nil {
+		return x.Locations
+	}
+	return nil
+}
+
 // if found, send the exact address
 // if not found, send the full list of existing brokers
 type LocateBrokerResponse_Resource struct {
@@ -3016,7 +3121,7 @@ type LocateBrokerResponse_Resource struct {
 func (x *LocateBrokerResponse_Resource) Reset() {
 	*x = LocateBrokerResponse_Resource{}
 	if protoimpl.UnsafeEnabled {
-		mi := &file_filer_proto_msgTypes[49]
+		mi := &file_filer_proto_msgTypes[50]
 		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
 		ms.StoreMessageInfo(mi)
 	}
@@ -3029,7 +3134,7 @@ func (x *LocateBrokerResponse_Resource) String() string {
 func (*LocateBrokerResponse_Resource) ProtoMessage() {}
 
 func (x *LocateBrokerResponse_Resource) ProtoReflect() protoreflect.Message {
-	mi := &file_filer_proto_msgTypes[49]
+	mi := &file_filer_proto_msgTypes[50]
 	if protoimpl.UnsafeEnabled && x != nil {
 		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
 		if ms.LoadMessageInfo() == nil {
@@ -3059,6 +3164,101 @@ func (x *LocateBrokerResponse_Resource) GetResourceCount() int32 {
 	return 0
 }
 
+type FilerConf_PathConf struct {
+	state         protoimpl.MessageState
+	sizeCache     protoimpl.SizeCache
+	unknownFields protoimpl.UnknownFields
+
+	LocationPrefix    string                      `protobuf:"bytes,1,opt,name=location_prefix,json=locationPrefix,proto3" json:"location_prefix,omitempty"`
+	Collection        string                      `protobuf:"bytes,2,opt,name=collection,proto3" json:"collection,omitempty"`
+	Replication       string                      `protobuf:"bytes,3,opt,name=replication,proto3" json:"replication,omitempty"`
+	Ttl               string                      `protobuf:"bytes,4,opt,name=ttl,proto3" json:"ttl,omitempty"`
+	DiskType          FilerConf_PathConf_DiskType `protobuf:"varint,5,opt,name=disk_type,json=diskType,proto3,enum=filer_pb.FilerConf_PathConf_DiskType" json:"disk_type,omitempty"`
+	Fsync             bool                        `protobuf:"varint,6,opt,name=fsync,proto3" json:"fsync,omitempty"`
+	VolumeGrowthCount uint32                      `protobuf:"varint,7,opt,name=volume_growth_count,json=volumeGrowthCount,proto3" json:"volume_growth_count,omitempty"`
+}
+
+func (x *FilerConf_PathConf) Reset() {
+	*x = FilerConf_PathConf{}
+	if protoimpl.UnsafeEnabled {
+		mi := &file_filer_proto_msgTypes[51]
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		ms.StoreMessageInfo(mi)
+	}
+}
+
+func (x *FilerConf_PathConf) String() string {
+	return protoimpl.X.MessageStringOf(x)
+}
+
+func (*FilerConf_PathConf) ProtoMessage() {}
+
+func (x *FilerConf_PathConf) ProtoReflect() protoreflect.Message {
+	mi := &file_filer_proto_msgTypes[51]
+	if protoimpl.UnsafeEnabled && x != nil {
+		ms := protoimpl.X.MessageStateOf(protoimpl.Pointer(x))
+		if ms.LoadMessageInfo() == nil {
+			ms.StoreMessageInfo(mi)
+		}
+		return ms
+	}
+	return mi.MessageOf(x)
+}
+
+// Deprecated: Use FilerConf_PathConf.ProtoReflect.Descriptor instead.
+func (*FilerConf_PathConf) Descriptor() ([]byte, []int) {
+	return file_filer_proto_rawDescGZIP(), []int{47, 0}
+}
+
+func (x *FilerConf_PathConf) GetLocationPrefix() string {
+	if x != nil {
+		return x.LocationPrefix
+	}
+	return ""
+}
+
+func (x *FilerConf_PathConf) GetCollection() string {
+	if x != nil {
+		return x.Collection
+	}
+	return ""
+}
+
+func (x *FilerConf_PathConf) GetReplication() string {
+	if x != nil {
+		return x.Replication
+	}
+	return ""
+}
+
+func (x *FilerConf_PathConf) GetTtl() string {
+	if x != nil {
+		return x.Ttl
+	}
+	return ""
+}
+
+func (x *FilerConf_PathConf) GetDiskType() FilerConf_PathConf_DiskType {
+	if x != nil {
+		return x.DiskType
+	}
+	return FilerConf_PathConf_NONE
+}
+
+func (x *FilerConf_PathConf) GetFsync() bool {
+	if x != nil {
+		return x.Fsync
+	}
+	return false
+}
+
+func (x *FilerConf_PathConf) GetVolumeGrowthCount() uint32 {
+	if x != nil {
+		return x.VolumeGrowthCount
+	}
+	return 0
+}
+
 var File_filer_proto protoreflect.FileDescriptor
 
 var file_filer_proto_rawDesc = []byte{
@@ -3433,7 +3633,33 @@ var file_filer_proto_rawDesc = []byte{
 	0x76, 0x61, 0x6c, 0x75, 0x65, 0x18, 0x02, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x05, 0x76, 0x61, 0x6c,
 	0x75, 0x65, 0x22, 0x25, 0x0a, 0x0d, 0x4b, 0x76, 0x50, 0x75, 0x74, 0x52, 0x65, 0x73, 0x70, 0x6f,
 	0x6e, 0x73, 0x65, 0x12, 0x14, 0x0a, 0x05, 0x65, 0x72, 0x72, 0x6f, 0x72, 0x18, 0x01, 0x20, 0x01,
-	0x28, 0x09, 0x52, 0x05, 0x65, 0x72, 0x72, 0x6f, 0x72, 0x32, 0xdc, 0x0c, 0x0a, 0x0c, 0x53, 0x65,
+	0x28, 0x09, 0x52, 0x05, 0x65, 0x72, 0x72, 0x6f, 0x72, 0x22, 0x9d, 0x03, 0x0a, 0x09, 0x46, 0x69,
+	0x6c, 0x65, 0x72, 0x43, 0x6f, 0x6e, 0x66, 0x12, 0x18, 0x0a, 0x07, 0x76, 0x65, 0x72, 0x73, 0x69,
+	0x6f, 0x6e, 0x18, 0x01, 0x20, 0x01, 0x28, 0x05, 0x52, 0x07, 0x76, 0x65, 0x72, 0x73, 0x69, 0x6f,
+	0x6e, 0x12, 0x3a, 0x0a, 0x09, 0x6c, 0x6f, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x73, 0x18, 0x02,
+	0x20, 0x03, 0x28, 0x0b, 0x32, 0x1c, 0x2e, 0x66, 0x69, 0x6c, 0x65, 0x72, 0x5f, 0x70, 0x62, 0x2e,
+	0x46, 0x69, 0x6c, 0x65, 0x72, 0x43, 0x6f, 0x6e, 0x66, 0x2e, 0x50, 0x61, 0x74, 0x68, 0x43, 0x6f,
+	0x6e, 0x66, 0x52, 0x09, 0x6c, 0x6f, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x73, 0x1a, 0xb9, 0x02,
+	0x0a, 0x08, 0x50, 0x61, 0x74, 0x68, 0x43, 0x6f, 0x6e, 0x66, 0x12, 0x27, 0x0a, 0x0f, 0x6c, 0x6f,
+	0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x5f, 0x70, 0x72, 0x65, 0x66, 0x69, 0x78, 0x18, 0x01, 0x20,
+	0x01, 0x28, 0x09, 0x52, 0x0e, 0x6c, 0x6f, 0x63, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0x50, 0x72, 0x65,
+	0x66, 0x69, 0x78, 0x12, 0x1e, 0x0a, 0x0a, 0x63, 0x6f, 0x6c, 0x6c, 0x65, 0x63, 0x74, 0x69, 0x6f,
+	0x6e, 0x18, 0x02, 0x20, 0x01, 0x28, 0x09, 0x52, 0x0a, 0x63, 0x6f, 0x6c, 0x6c, 0x65, 0x63, 0x74,
+	0x69, 0x6f, 0x6e, 0x12, 0x20, 0x0a, 0x0b, 0x72, 0x65, 0x70, 0x6c, 0x69, 0x63, 0x61, 0x74, 0x69,
+	0x6f, 0x6e, 0x18, 0x03, 0x20, 0x01, 0x28, 0x09, 0x52, 0x0b, 0x72, 0x65, 0x70, 0x6c, 0x69, 0x63,
+	0x61, 0x74, 0x69, 0x6f, 0x6e, 0x12, 0x10, 0x0a, 0x03, 0x74, 0x74, 0x6c, 0x18, 0x04, 0x20, 0x01,
+	0x28, 0x09, 0x52, 0x03, 0x74, 0x74, 0x6c, 0x12, 0x42, 0x0a, 0x09, 0x64, 0x69, 0x73, 0x6b, 0x5f,
+	0x74, 0x79, 0x70, 0x65, 0x18, 0x05, 0x20, 0x01, 0x28, 0x0e, 0x32, 0x25, 0x2e, 0x66, 0x69, 0x6c,
+	0x65, 0x72, 0x5f, 0x70, 0x62, 0x2e, 0x46, 0x69, 0x6c, 0x65, 0x72, 0x43, 0x6f, 0x6e, 0x66, 0x2e,
+	0x50, 0x61, 0x74, 0x68, 0x43, 0x6f, 0x6e, 0x66, 0x2e, 0x44, 0x69, 0x73, 0x6b, 0x54, 0x79, 0x70,
+	0x65, 0x52, 0x08, 0x64, 0x69, 0x73, 0x6b, 0x54, 0x79, 0x70, 0x65, 0x12, 0x14, 0x0a, 0x05, 0x66,
+	0x73, 0x79, 0x6e, 0x63, 0x18, 0x06, 0x20, 0x01, 0x28, 0x08, 0x52, 0x05, 0x66, 0x73, 0x79, 0x6e,
+	0x63, 0x12, 0x2e, 0x0a, 0x13, 0x76, 0x6f, 0x6c, 0x75, 0x6d, 0x65, 0x5f, 0x67, 0x72, 0x6f, 0x77,
+	0x74, 0x68, 0x5f, 0x63, 0x6f, 0x75, 0x6e, 0x74, 0x18, 0x07, 0x20, 0x01, 0x28, 0x0d, 0x52, 0x11,
+	0x76, 0x6f, 0x6c, 0x75, 0x6d, 0x65, 0x47, 0x72, 0x6f, 0x77, 0x74, 0x68, 0x43, 0x6f, 0x75, 0x6e,
+	0x74, 0x22, 0x26, 0x0a, 0x08, 0x44, 0x69, 0x73, 0x6b, 0x54, 0x79, 0x70, 0x65, 0x12, 0x08, 0x0a,
+	0x04, 0x4e, 0x4f, 0x4e, 0x45, 0x10, 0x00, 0x12, 0x07, 0x0a, 0x03, 0x48, 0x44, 0x44, 0x10, 0x01,
+	0x12, 0x07, 0x0a, 0x03, 0x53, 0x53, 0x44, 0x10, 0x02, 0x32, 0xdc, 0x0c, 0x0a, 0x0c, 0x53, 0x65,
 	0x61, 0x77, 0x65, 0x65, 0x64, 0x46, 0x69, 0x6c, 0x65, 0x72, 0x12, 0x67, 0x0a, 0x14, 0x4c, 0x6f,
 	0x6f, 0x6b, 0x75, 0x70, 0x44, 0x69, 0x72, 0x65, 0x63, 0x74, 0x6f, 0x72, 0x79, 0x45, 0x6e, 0x74,
 	0x72, 0x79, 0x12, 0x25, 0x2e, 0x66, 0x69, 0x6c, 0x65, 0x72, 0x5f, 0x70, 0x62, 0x2e, 0x4c, 0x6f,
@@ -3556,123 +3782,129 @@ func file_filer_proto_rawDescGZIP() []byte {
 	return file_filer_proto_rawDescData
 }
 
-var file_filer_proto_msgTypes = make([]protoimpl.MessageInfo, 50)
+var file_filer_proto_enumTypes = make([]protoimpl.EnumInfo, 1)
+var file_filer_proto_msgTypes = make([]protoimpl.MessageInfo, 52)
 var file_filer_proto_goTypes = []interface{}{
-	(*LookupDirectoryEntryRequest)(nil),   // 0: filer_pb.LookupDirectoryEntryRequest
+	(FilerConf_PathConf_DiskType)(0),      // 0: filer_pb.FilerConf.PathConf.DiskType
-	(*LookupDirectoryEntryResponse)(nil),  // 1: filer_pb.LookupDirectoryEntryResponse
+	(*LookupDirectoryEntryRequest)(nil),   // 1: filer_pb.LookupDirectoryEntryRequest
-	(*ListEntriesRequest)(nil),            // 2: filer_pb.ListEntriesRequest
+	(*LookupDirectoryEntryResponse)(nil),  // 2: filer_pb.LookupDirectoryEntryResponse
-	(*ListEntriesResponse)(nil),           // 3: filer_pb.ListEntriesResponse
+	(*ListEntriesRequest)(nil),            // 3: filer_pb.ListEntriesRequest
-	(*Entry)(nil),                         // 4: filer_pb.Entry
+	(*ListEntriesResponse)(nil),           // 4: filer_pb.ListEntriesResponse
-	(*FullEntry)(nil),                     // 5: filer_pb.FullEntry
+	(*Entry)(nil),                         // 5: filer_pb.Entry
-	(*EventNotification)(nil),             // 6: filer_pb.EventNotification
+	(*FullEntry)(nil),                     // 6: filer_pb.FullEntry
-	(*FileChunk)(nil),                     // 7: filer_pb.FileChunk
+	(*EventNotification)(nil),             // 7: filer_pb.EventNotification
-	(*FileChunkManifest)(nil),             // 8: filer_pb.FileChunkManifest
+	(*FileChunk)(nil),                     // 8: filer_pb.FileChunk
-	(*FileId)(nil),                        // 9: filer_pb.FileId
+	(*FileChunkManifest)(nil),             // 9: filer_pb.FileChunkManifest
-	(*FuseAttributes)(nil),                // 10: filer_pb.FuseAttributes
+	(*FileId)(nil),                        // 10: filer_pb.FileId
-	(*CreateEntryRequest)(nil),            // 11: filer_pb.CreateEntryRequest
+	(*FuseAttributes)(nil),                // 11: filer_pb.FuseAttributes
-	(*CreateEntryResponse)(nil),           // 12: filer_pb.CreateEntryResponse
+	(*CreateEntryRequest)(nil),            // 12: filer_pb.CreateEntryRequest
-	(*UpdateEntryRequest)(nil),            // 13: filer_pb.UpdateEntryRequest
+	(*CreateEntryResponse)(nil),           // 13: filer_pb.CreateEntryResponse
-	(*UpdateEntryResponse)(nil),           // 14: filer_pb.UpdateEntryResponse
+	(*UpdateEntryRequest)(nil),            // 14: filer_pb.UpdateEntryRequest
-	(*AppendToEntryRequest)(nil),          // 15: filer_pb.AppendToEntryRequest
+	(*UpdateEntryResponse)(nil),           // 15: filer_pb.UpdateEntryResponse
-	(*AppendToEntryResponse)(nil),         // 16: filer_pb.AppendToEntryResponse
+	(*AppendToEntryRequest)(nil),          // 16: filer_pb.AppendToEntryRequest
-	(*DeleteEntryRequest)(nil),            // 17: filer_pb.DeleteEntryRequest
+	(*AppendToEntryResponse)(nil),         // 17: filer_pb.AppendToEntryResponse
-	(*DeleteEntryResponse)(nil),           // 18: filer_pb.DeleteEntryResponse
+	(*DeleteEntryRequest)(nil),            // 18: filer_pb.DeleteEntryRequest
-	(*AtomicRenameEntryRequest)(nil),      // 19: filer_pb.AtomicRenameEntryRequest
+	(*DeleteEntryResponse)(nil),           // 19: filer_pb.DeleteEntryResponse
-	(*AtomicRenameEntryResponse)(nil),     // 20: filer_pb.AtomicRenameEntryResponse
+	(*AtomicRenameEntryRequest)(nil),      // 20: filer_pb.AtomicRenameEntryRequest
-	(*AssignVolumeRequest)(nil),           // 21: filer_pb.AssignVolumeRequest
+	(*AtomicRenameEntryResponse)(nil),     // 21: filer_pb.AtomicRenameEntryResponse
-	(*AssignVolumeResponse)(nil),          // 22: filer_pb.AssignVolumeResponse
+	(*AssignVolumeRequest)(nil),           // 22: filer_pb.AssignVolumeRequest
-	(*LookupVolumeRequest)(nil),           // 23: filer_pb.LookupVolumeRequest
+	(*AssignVolumeResponse)(nil),          // 23: filer_pb.AssignVolumeResponse
-	(*Locations)(nil),                     // 24: filer_pb.Locations
+	(*LookupVolumeRequest)(nil),           // 24: filer_pb.LookupVolumeRequest
-	(*Location)(nil),                      // 25: filer_pb.Location
+	(*Locations)(nil),                     // 25: filer_pb.Locations
-	(*LookupVolumeResponse)(nil),          // 26: filer_pb.LookupVolumeResponse
+	(*Location)(nil),                      // 26: filer_pb.Location
-	(*Collection)(nil),                    // 27: filer_pb.Collection
+	(*LookupVolumeResponse)(nil),          // 27: filer_pb.LookupVolumeResponse
-	(*CollectionListRequest)(nil),         // 28: filer_pb.CollectionListRequest
+	(*Collection)(nil),                    // 28: filer_pb.Collection
-	(*CollectionListResponse)(nil),        // 29: filer_pb.CollectionListResponse
+	(*CollectionListRequest)(nil),         // 29: filer_pb.CollectionListRequest
-	(*DeleteCollectionRequest)(nil),       // 30: filer_pb.DeleteCollectionRequest
+	(*CollectionListResponse)(nil),        // 30: filer_pb.CollectionListResponse
-	(*DeleteCollectionResponse)(nil),      // 31: filer_pb.DeleteCollectionResponse
+	(*DeleteCollectionRequest)(nil),       // 31: filer_pb.DeleteCollectionRequest
-	(*StatisticsRequest)(nil),             // 32: filer_pb.StatisticsRequest
+	(*DeleteCollectionResponse)(nil),      // 32: filer_pb.DeleteCollectionResponse
-	(*StatisticsResponse)(nil),            // 33: filer_pb.StatisticsResponse
+	(*StatisticsRequest)(nil),             // 33: filer_pb.StatisticsRequest
-	(*GetFilerConfigurationRequest)(nil),  // 34: filer_pb.GetFilerConfigurationRequest
+	(*StatisticsResponse)(nil),            // 34: filer_pb.StatisticsResponse
-	(*GetFilerConfigurationResponse)(nil), // 35: filer_pb.GetFilerConfigurationResponse
+	(*GetFilerConfigurationRequest)(nil),  // 35: filer_pb.GetFilerConfigurationRequest
-	(*SubscribeMetadataRequest)(nil),      // 36: filer_pb.SubscribeMetadataRequest
+	(*GetFilerConfigurationResponse)(nil), // 36: filer_pb.GetFilerConfigurationResponse
-	(*SubscribeMetadataResponse)(nil),     // 37: filer_pb.SubscribeMetadataResponse
+	(*SubscribeMetadataRequest)(nil),      // 37: filer_pb.SubscribeMetadataRequest
-	(*LogEntry)(nil),                      // 38: filer_pb.LogEntry
+	(*SubscribeMetadataResponse)(nil),     // 38: filer_pb.SubscribeMetadataResponse
-	(*KeepConnectedRequest)(nil),          // 39: filer_pb.KeepConnectedRequest
+	(*LogEntry)(nil),                      // 39: filer_pb.LogEntry
-	(*KeepConnectedResponse)(nil),         // 40: filer_pb.KeepConnectedResponse
+	(*KeepConnectedRequest)(nil),          // 40: filer_pb.KeepConnectedRequest
-	(*LocateBrokerRequest)(nil),           // 41: filer_pb.LocateBrokerRequest
+	(*KeepConnectedResponse)(nil),         // 41: filer_pb.KeepConnectedResponse
-	(*LocateBrokerResponse)(nil),          // 42: filer_pb.LocateBrokerResponse
+	(*LocateBrokerRequest)(nil),           // 42: filer_pb.LocateBrokerRequest
-	(*KvGetRequest)(nil),                  // 43: filer_pb.KvGetRequest
+	(*LocateBrokerResponse)(nil),          // 43: filer_pb.LocateBrokerResponse
-	(*KvGetResponse)(nil),                 // 44: filer_pb.KvGetResponse
+	(*KvGetRequest)(nil),                  // 44: filer_pb.KvGetRequest
-	(*KvPutRequest)(nil),                  // 45: filer_pb.KvPutRequest
+	(*KvGetResponse)(nil),                 // 45: filer_pb.KvGetResponse
-	(*KvPutResponse)(nil),                 // 46: filer_pb.KvPutResponse
+	(*KvPutRequest)(nil),                  // 46: filer_pb.KvPutRequest
-	nil,                                   // 47: filer_pb.Entry.ExtendedEntry
+	(*KvPutResponse)(nil),                 // 47: filer_pb.KvPutResponse
-	nil,                                   // 48: filer_pb.LookupVolumeResponse.LocationsMapEntry
+	(*FilerConf)(nil),                     // 48: filer_pb.FilerConf
-	(*LocateBrokerResponse_Resource)(nil), // 49: filer_pb.LocateBrokerResponse.Resource
+	nil,                                   // 49: filer_pb.Entry.ExtendedEntry
+	nil,                                   // 50: filer_pb.LookupVolumeResponse.LocationsMapEntry
+	(*LocateBrokerResponse_Resource)(nil), // 51: filer_pb.LocateBrokerResponse.Resource
+	(*FilerConf_PathConf)(nil),            // 52: filer_pb.FilerConf.PathConf
 }
 var file_filer_proto_depIdxs = []int32{
-	4,  // 0: filer_pb.LookupDirectoryEntryResponse.entry:type_name -> filer_pb.Entry
+	5,  // 0: filer_pb.LookupDirectoryEntryResponse.entry:type_name -> filer_pb.Entry
-	4,  // 1: filer_pb.ListEntriesResponse.entry:type_name -> filer_pb.Entry
+	5,  // 1: filer_pb.ListEntriesResponse.entry:type_name -> filer_pb.Entry
-	7,  // 2: filer_pb.Entry.chunks:type_name -> filer_pb.FileChunk
+	8,  // 2: filer_pb.Entry.chunks:type_name -> filer_pb.FileChunk
-	10, // 3: filer_pb.Entry.attributes:type_name -> filer_pb.FuseAttributes
+	11, // 3: filer_pb.Entry.attributes:type_name -> filer_pb.FuseAttributes
-	47, // 4: filer_pb.Entry.extended:type_name -> filer_pb.Entry.ExtendedEntry
+	49, // 4: filer_pb.Entry.extended:type_name -> filer_pb.Entry.ExtendedEntry
-	4,  // 5: filer_pb.FullEntry.entry:type_name -> filer_pb.Entry
+	5,  // 5: filer_pb.FullEntry.entry:type_name -> filer_pb.Entry
-	4,  // 6: filer_pb.EventNotification.old_entry:type_name -> filer_pb.Entry
+	5,  // 6: filer_pb.EventNotification.old_entry:type_name -> filer_pb.Entry
-	4,  // 7: filer_pb.EventNotification.new_entry:type_name -> filer_pb.Entry
+	5,  // 7: filer_pb.EventNotification.new_entry:type_name -> filer_pb.Entry
-	9,  // 8: filer_pb.FileChunk.fid:type_name -> filer_pb.FileId
+	10, // 8: filer_pb.FileChunk.fid:type_name -> filer_pb.FileId
-	9,  // 9: filer_pb.FileChunk.source_fid:type_name -> filer_pb.FileId
+	10, // 9: filer_pb.FileChunk.source_fid:type_name -> filer_pb.FileId
-	7,  // 10: filer_pb.FileChunkManifest.chunks:type_name -> filer_pb.FileChunk
+	8,  // 10: filer_pb.FileChunkManifest.chunks:type_name -> filer_pb.FileChunk
-	4,  // 11: filer_pb.CreateEntryRequest.entry:type_name -> filer_pb.Entry
+	5,  // 11: filer_pb.CreateEntryRequest.entry:type_name -> filer_pb.Entry
-	4,  // 12: filer_pb.UpdateEntryRequest.entry:type_name -> filer_pb.Entry
+	5,  // 12: filer_pb.UpdateEntryRequest.entry:type_name -> filer_pb.Entry
-	7,  // 13: filer_pb.AppendToEntryRequest.chunks:type_name -> filer_pb.FileChunk
+	8,  // 13: filer_pb.AppendToEntryRequest.chunks:type_name -> filer_pb.FileChunk
-	25, // 14: filer_pb.Locations.locations:type_name -> filer_pb.Location
+	26, // 14: filer_pb.Locations.locations:type_name -> filer_pb.Location
-	48, // 15: filer_pb.LookupVolumeResponse.locations_map:type_name -> filer_pb.LookupVolumeResponse.LocationsMapEntry
+	50, // 15: filer_pb.LookupVolumeResponse.locations_map:type_name -> filer_pb.LookupVolumeResponse.LocationsMapEntry
-	27, // 16: filer_pb.CollectionListResponse.collections:type_name -> filer_pb.Collection
+	28, // 16: filer_pb.CollectionListResponse.collections:type_name -> filer_pb.Collection
-	6,  // 17: filer_pb.SubscribeMetadataResponse.event_notification:type_name -> filer_pb.EventNotification
+	7,  // 17: filer_pb.SubscribeMetadataResponse.event_notification:type_name -> filer_pb.EventNotification
-	49, // 18: filer_pb.LocateBrokerResponse.resources:type_name -> filer_pb.LocateBrokerResponse.Resource
+	51, // 18: filer_pb.LocateBrokerResponse.resources:type_name -> filer_pb.LocateBrokerResponse.Resource
-	24, // 19: filer_pb.LookupVolumeResponse.LocationsMapEntry.value:type_name -> filer_pb.Locations
+	52, // 19: filer_pb.FilerConf.locations:type_name -> filer_pb.FilerConf.PathConf
-	0,  // 20: filer_pb.SeaweedFiler.LookupDirectoryEntry:input_type -> filer_pb.LookupDirectoryEntryRequest
+	25, // 20: filer_pb.LookupVolumeResponse.LocationsMapEntry.value:type_name -> filer_pb.Locations
-	2,  // 21: filer_pb.SeaweedFiler.ListEntries:input_type -> filer_pb.ListEntriesRequest
+	0,  // 21: filer_pb.FilerConf.PathConf.disk_type:type_name -> filer_pb.FilerConf.PathConf.DiskType
-	11, // 22: filer_pb.SeaweedFiler.CreateEntry:input_type -> filer_pb.CreateEntryRequest
+	1,  // 22: filer_pb.SeaweedFiler.LookupDirectoryEntry:input_type -> filer_pb.LookupDirectoryEntryRequest
-	13, // 23: filer_pb.SeaweedFiler.UpdateEntry:input_type -> filer_pb.UpdateEntryRequest
+	3,  // 23: filer_pb.SeaweedFiler.ListEntries:input_type -> filer_pb.ListEntriesRequest
-	15, // 24: filer_pb.SeaweedFiler.AppendToEntry:input_type -> filer_pb.AppendToEntryRequest
+	12, // 24: filer_pb.SeaweedFiler.CreateEntry:input_type -> filer_pb.CreateEntryRequest
-	17, // 25: filer_pb.SeaweedFiler.DeleteEntry:input_type -> filer_pb.DeleteEntryRequest
+	14, // 25: filer_pb.SeaweedFiler.UpdateEntry:input_type -> filer_pb.UpdateEntryRequest
-	19, // 26: filer_pb.SeaweedFiler.AtomicRenameEntry:input_type -> filer_pb.AtomicRenameEntryRequest
+	16, // 26: filer_pb.SeaweedFiler.AppendToEntry:input_type -> filer_pb.AppendToEntryRequest
-	21, // 27: filer_pb.SeaweedFiler.AssignVolume:input_type -> filer_pb.AssignVolumeRequest
+	18, // 27: filer_pb.SeaweedFiler.DeleteEntry:input_type -> filer_pb.DeleteEntryRequest
-	23, // 28: filer_pb.SeaweedFiler.LookupVolume:input_type -> filer_pb.LookupVolumeRequest
+	20, // 28: filer_pb.SeaweedFiler.AtomicRenameEntry:input_type -> filer_pb.AtomicRenameEntryRequest
-	28, // 29: filer_pb.SeaweedFiler.CollectionList:input_type -> filer_pb.CollectionListRequest
+	22, // 29: filer_pb.SeaweedFiler.AssignVolume:input_type -> filer_pb.AssignVolumeRequest
-	30, // 30: filer_pb.SeaweedFiler.DeleteCollection:input_type -> filer_pb.DeleteCollectionRequest
+	24, // 30: filer_pb.SeaweedFiler.LookupVolume:input_type -> filer_pb.LookupVolumeRequest
-	32, // 31: filer_pb.SeaweedFiler.Statistics:input_type -> filer_pb.StatisticsRequest
+	29, // 31: filer_pb.SeaweedFiler.CollectionList:input_type -> filer_pb.CollectionListRequest
-	34, // 32: filer_pb.SeaweedFiler.GetFilerConfiguration:input_type -> filer_pb.GetFilerConfigurationRequest
+	31, // 32: filer_pb.SeaweedFiler.DeleteCollection:input_type -> filer_pb.DeleteCollectionRequest
-	36, // 33: filer_pb.SeaweedFiler.SubscribeMetadata:input_type -> filer_pb.SubscribeMetadataRequest
+	33, // 33: filer_pb.SeaweedFiler.Statistics:input_type -> filer_pb.StatisticsRequest
-	36, // 34: filer_pb.SeaweedFiler.SubscribeLocalMetadata:input_type -> filer_pb.SubscribeMetadataRequest
+	35, // 34: filer_pb.SeaweedFiler.GetFilerConfiguration:input_type -> filer_pb.GetFilerConfigurationRequest
-	39, // 35: filer_pb.SeaweedFiler.KeepConnected:input_type -> filer_pb.KeepConnectedRequest
+	37, // 35: filer_pb.SeaweedFiler.SubscribeMetadata:input_type -> filer_pb.SubscribeMetadataRequest
-	41, // 36: filer_pb.SeaweedFiler.LocateBroker:input_type -> filer_pb.LocateBrokerRequest
+	37, // 36: filer_pb.SeaweedFiler.SubscribeLocalMetadata:input_type -> filer_pb.SubscribeMetadataRequest
-	43, // 37: filer_pb.SeaweedFiler.KvGet:input_type -> filer_pb.KvGetRequest
+	40, // 37: filer_pb.SeaweedFiler.KeepConnected:input_type -> filer_pb.KeepConnectedRequest
-	45, // 38: filer_pb.SeaweedFiler.KvPut:input_type -> filer_pb.KvPutRequest
+	42, // 38: filer_pb.SeaweedFiler.LocateBroker:input_type -> filer_pb.LocateBrokerRequest
-	1,  // 39: filer_pb.SeaweedFiler.LookupDirectoryEntry:output_type -> filer_pb.LookupDirectoryEntryResponse
+	44, // 39: filer_pb.SeaweedFiler.KvGet:input_type -> filer_pb.KvGetRequest
-	3,  // 40: filer_pb.SeaweedFiler.ListEntries:output_type -> filer_pb.ListEntriesResponse
+	46, // 40: filer_pb.SeaweedFiler.KvPut:input_type -> filer_pb.KvPutRequest
-	12, // 41: filer_pb.SeaweedFiler.CreateEntry:output_type -> filer_pb.CreateEntryResponse
+	2,  // 41: filer_pb.SeaweedFiler.LookupDirectoryEntry:output_type -> filer_pb.LookupDirectoryEntryResponse
-	14, // 42: filer_pb.SeaweedFiler.UpdateEntry:output_type -> filer_pb.UpdateEntryResponse
+	4,  // 42: filer_pb.SeaweedFiler.ListEntries:output_type -> filer_pb.ListEntriesResponse
-	16, // 43: filer_pb.SeaweedFiler.AppendToEntry:output_type -> filer_pb.AppendToEntryResponse
+	13, // 43: filer_pb.SeaweedFiler.CreateEntry:output_type -> filer_pb.CreateEntryResponse
-	18, // 44: filer_pb.SeaweedFiler.DeleteEntry:output_type -> filer_pb.DeleteEntryResponse
+	15, // 44: filer_pb.SeaweedFiler.UpdateEntry:output_type -> filer_pb.UpdateEntryResponse
-	20, // 45: filer_pb.SeaweedFiler.AtomicRenameEntry:output_type -> filer_pb.AtomicRenameEntryResponse
+	17, // 45: filer_pb.SeaweedFiler.AppendToEntry:output_type -> filer_pb.AppendToEntryResponse
-	22, // 46: filer_pb.SeaweedFiler.AssignVolume:output_type -> filer_pb.AssignVolumeResponse
+	19, // 46: filer_pb.SeaweedFiler.DeleteEntry:output_type -> filer_pb.DeleteEntryResponse
-	26, // 47: filer_pb.SeaweedFiler.LookupVolume:output_type -> filer_pb.LookupVolumeResponse
+	21, // 47: filer_pb.SeaweedFiler.AtomicRenameEntry:output_type -> filer_pb.AtomicRenameEntryResponse
-	29, // 48: filer_pb.SeaweedFiler.CollectionList:output_type -> filer_pb.CollectionListResponse
+	23, // 48: filer_pb.SeaweedFiler.AssignVolume:output_type -> filer_pb.AssignVolumeResponse
-	31, // 49: filer_pb.SeaweedFiler.DeleteCollection:output_type -> filer_pb.DeleteCollectionResponse
+	27, // 49: filer_pb.SeaweedFiler.LookupVolume:output_type -> filer_pb.LookupVolumeResponse
-	33, // 50: filer_pb.SeaweedFiler.Statistics:output_type -> filer_pb.StatisticsResponse
+	30, // 50: filer_pb.SeaweedFiler.CollectionList:output_type -> filer_pb.CollectionListResponse
-	35, // 51: filer_pb.SeaweedFiler.GetFilerConfiguration:output_type -> filer_pb.GetFilerConfigurationResponse
+	32, // 51: filer_pb.SeaweedFiler.DeleteCollection:output_type -> filer_pb.DeleteCollectionResponse
-	37, // 52: filer_pb.SeaweedFiler.SubscribeMetadata:output_type -> filer_pb.SubscribeMetadataResponse
+	34, // 52: filer_pb.SeaweedFiler.Statistics:output_type -> filer_pb.StatisticsResponse
-	37, // 53: filer_pb.SeaweedFiler.SubscribeLocalMetadata:output_type -> filer_pb.SubscribeMetadataResponse
+	36, // 53: filer_pb.SeaweedFiler.GetFilerConfiguration:output_type -> filer_pb.GetFilerConfigurationResponse
-	40, // 54: filer_pb.SeaweedFiler.KeepConnected:output_type -> filer_pb.KeepConnectedResponse
+	38, // 54: filer_pb.SeaweedFiler.SubscribeMetadata:output_type -> filer_pb.SubscribeMetadataResponse
-	42, // 55: filer_pb.SeaweedFiler.LocateBroker:output_type -> filer_pb.LocateBrokerResponse
+	38, // 55: filer_pb.SeaweedFiler.SubscribeLocalMetadata:output_type -> filer_pb.SubscribeMetadataResponse
-	44, // 56: filer_pb.SeaweedFiler.KvGet:output_type -> filer_pb.KvGetResponse
+	41, // 56: filer_pb.SeaweedFiler.KeepConnected:output_type -> filer_pb.KeepConnectedResponse
-	46, // 57: filer_pb.SeaweedFiler.KvPut:output_type -> filer_pb.KvPutResponse
+	43, // 57: filer_pb.SeaweedFiler.LocateBroker:output_type -> filer_pb.LocateBrokerResponse
-	39, // [39:58] is the sub-list for method output_type
+	45, // 58: filer_pb.SeaweedFiler.KvGet:output_type -> filer_pb.KvGetResponse
-	20, // [20:39] is the sub-list for method input_type
+	47, // 59: filer_pb.SeaweedFiler.KvPut:output_type -> filer_pb.KvPutResponse
-	20, // [20:20] is the sub-list for extension type_name
+	41, // [41:60] is the sub-list for method output_type
-	20, // [20:20] is the sub-list for extension extendee
+	22, // [22:41] is the sub-list for method input_type
-	0,  // [0:20] is the sub-list for field type_name
+	22, // [22:22] is the sub-list for extension type_name
+	22, // [22:22] is the sub-list for extension extendee
+	0,  // [0:22] is the sub-list for field type_name
 }
 
 func init() { file_filer_proto_init() }
@@ -4245,7 +4477,19 @@ func file_filer_proto_init() {
 				return nil
 			}
 		}
-		file_filer_proto_msgTypes[49].Exporter = func(v interface{}, i int) interface{} {
+		file_filer_proto_msgTypes[47].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*FilerConf); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
+		file_filer_proto_msgTypes[50].Exporter = func(v interface{}, i int) interface{} {
 			switch v := v.(*LocateBrokerResponse_Resource); i {
 			case 0:
 				return &v.state
@@ -4257,19 +4501,32 @@ func file_filer_proto_init() {
 				return nil
 			}
 		}
+		file_filer_proto_msgTypes[51].Exporter = func(v interface{}, i int) interface{} {
+			switch v := v.(*FilerConf_PathConf); i {
+			case 0:
+				return &v.state
+			case 1:
+				return &v.sizeCache
+			case 2:
+				return &v.unknownFields
+			default:
+				return nil
+			}
+		}
 	}
 	type x struct{}
 	out := protoimpl.TypeBuilder{
 		File: protoimpl.DescBuilder{
 			GoPackagePath: reflect.TypeOf(x{}).PkgPath(),
 			RawDescriptor: file_filer_proto_rawDesc,
-			NumEnums:      0,
+			NumEnums:      1,
-			NumMessages:   50,
+			NumMessages:   52,
 			NumExtensions: 0,
 			NumServices:   1,
 		},
 		GoTypes:           file_filer_proto_goTypes,
 		DependencyIndexes: file_filer_proto_depIdxs,
+		EnumInfos:         file_filer_proto_enumTypes,
 		MessageInfos:      file_filer_proto_msgTypes,
 	}.Build()
 	File_filer_proto = out.File

weed/pb/filer_pb/filer_pb_helper.go

@@ -8,6 +8,8 @@ import (
 
 	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/storage/needle"
+	"github.com/golang/protobuf/proto"
+	"github.com/viant/ptrie"
 )
 
 func ToFileIdObject(fileIdStr string) (*FileId, error) {
@@ -121,3 +123,27 @@ func LookupEntry(client SeaweedFilerClient, request *LookupDirectoryEntryRequest
 }
 
 var ErrNotFound = errors.New("filer: no entry is found in filer store")
+
+func IsCreate(event *SubscribeMetadataResponse) bool {
+	return event.EventNotification.NewEntry != nil && event.EventNotification.OldEntry == nil
+}
+func IsUpdate(event *SubscribeMetadataResponse) bool {
+	return event.EventNotification.NewEntry != nil &&
+		event.EventNotification.OldEntry != nil &&
+		event.Directory == event.EventNotification.NewParentPath
+}
+func IsDelete(event *SubscribeMetadataResponse) bool {
+	return event.EventNotification.NewEntry == nil && event.EventNotification.OldEntry != nil
+}
+func IsRename(event *SubscribeMetadataResponse) bool {
+	return event.EventNotification.NewEntry != nil &&
+		event.EventNotification.OldEntry != nil &&
+		event.Directory != event.EventNotification.NewParentPath
+}
+
+var _ = ptrie.KeyProvider(&FilerConf_PathConf{})
+
+func (fp *FilerConf_PathConf) Key() interface{} {
+	key, _ := proto.Marshal(fp)
+	return string(key)
+}

weed/pb/master.proto

@@ -130,6 +130,7 @@ message VolumeLocation {
     repeated uint32 new_vids = 3;
     repeated uint32 deleted_vids = 4;
     string leader = 5; // optional when leader is not itself
+    string data_center = 6; // optional when DataCenter is in use
 }
 
 message LookupVolumeRequest {
@@ -187,11 +188,6 @@ message StatisticsResponse {
 //
 // collection related
 //
-
-message StorageType {
-    string replication = 1;
-    string ttl = 2;
-}
 message Collection {
     string name = 1;
 }

weed/s3api/auth_credentials.go

@@ -3,10 +3,11 @@ package s3api
 import (
 	"bytes"
 	"fmt"
-	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
 	"io/ioutil"
 	"net/http"
 
+	xhttp "github.com/chrislusf/seaweedfs/weed/s3api/http"
+	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
 	"github.com/golang/protobuf/jsonpb"
 
 	"github.com/chrislusf/seaweedfs/weed/glog"
@@ -128,8 +129,14 @@ func (iam *IdentityAccessManagement) Auth(f http.HandlerFunc, action Action) htt
 	}
 
 	return func(w http.ResponseWriter, r *http.Request) {
-		errCode := iam.authRequest(r, action)
+		identity, errCode := iam.authRequest(r, action)
 		if errCode == s3err.ErrNone {
+			if identity != nil && identity.Name != "" {
+				r.Header.Set(xhttp.AmzIdentityId, identity.Name)
+				if identity.isAdmin() {
+					r.Header.Set(xhttp.AmzIsAdmin, "true")
+				}
+			}
 			f(w, r)
 			return
 		}
@@ -138,16 +145,16 @@ func (iam *IdentityAccessManagement) Auth(f http.HandlerFunc, action Action) htt
 }
 
 // check whether the request has valid access keys
-func (iam *IdentityAccessManagement) authRequest(r *http.Request, action Action) s3err.ErrorCode {
+func (iam *IdentityAccessManagement) authRequest(r *http.Request, action Action) (*Identity, s3err.ErrorCode) {
 	var identity *Identity
 	var s3Err s3err.ErrorCode
 	var found bool
 	switch getRequestAuthType(r) {
 	case authTypeStreamingSigned:
-		return s3err.ErrNone
+		return identity, s3err.ErrNone
 	case authTypeUnknown:
 		glog.V(3).Infof("unknown auth type")
-		return s3err.ErrAccessDenied
+		return identity, s3err.ErrAccessDenied
 	case authTypePresignedV2, authTypeSignedV2:
 		glog.V(3).Infof("v2 auth type")
 		identity, s3Err = iam.isReqAuthenticatedV2(r)
@@ -156,22 +163,22 @@ func (iam *IdentityAccessManagement) authRequest(r *http.Request, action Action)
 		identity, s3Err = iam.reqSignatureV4Verify(r)
 	case authTypePostPolicy:
 		glog.V(3).Infof("post policy auth type")
-		return s3err.ErrNone
+		return identity, s3err.ErrNone
 	case authTypeJWT:
 		glog.V(3).Infof("jwt auth type")
-		return s3err.ErrNotImplemented
+		return identity, s3err.ErrNotImplemented
 	case authTypeAnonymous:
 		identity, found = iam.lookupAnonymous()
 		if !found {
-			return s3err.ErrAccessDenied
+			return identity, s3err.ErrAccessDenied
 		}
 	default:
-		return s3err.ErrNotImplemented
+		return identity, s3err.ErrNotImplemented
 	}
 
 	glog.V(3).Infof("auth error: %v", s3Err)
 	if s3Err != s3err.ErrNone {
-		return s3Err
+		return identity, s3Err
 	}
 
 	glog.V(3).Infof("user name: %v actions: %v", identity.Name, identity.Actions)
@@ -179,19 +186,17 @@ func (iam *IdentityAccessManagement) authRequest(r *http.Request, action Action)
 	bucket, _ := getBucketAndObject(r)
 
 	if !identity.canDo(action, bucket) {
-		return s3err.ErrAccessDenied
+		return identity, s3err.ErrAccessDenied
 	}
 
-	return s3err.ErrNone
+	return identity, s3err.ErrNone
 
 }
 
 func (identity *Identity) canDo(action Action, bucket string) bool {
-	for _, a := range identity.Actions {
+	if identity.isAdmin() {
-		if a == "Admin" {
 		return true
 	}
-	}
 	for _, a := range identity.Actions {
 		if a == action {
 			return true
@@ -208,3 +213,12 @@ func (identity *Identity) canDo(action Action, bucket string) bool {
 	}
 	return false
 }
+
+func (identity *Identity) isAdmin() bool {
+	for _, a := range identity.Actions {
+		if a == "Admin" {
+			return true
+		}
+	}
+	return false
+}

weed/s3api/filer_util.go

@@ -3,13 +3,14 @@ package s3api
 import (
 	"context"
 	"fmt"
+	"strings"
+
 	"github.com/chrislusf/seaweedfs/weed/glog"
-	"github.com/chrislusf/seaweedfs/weed/pb"
 	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
 	"github.com/chrislusf/seaweedfs/weed/pb/iam_pb"
 	proto "github.com/golang/protobuf/proto"
 	"google.golang.org/grpc"
-	"strings"
+	"github.com/chrislusf/seaweedfs/weed/util"
 )
 
 func (s3a *S3ApiServer) mkdir(parentDirectoryPath string, dirName string, fn func(entry *filer_pb.Entry)) error {
@@ -78,6 +79,11 @@ func (s3a *S3ApiServer) exists(parentDirectoryPath string, entryName string, isD
 
 }
 
+func (s3a *S3ApiServer) getEntry(parentDirectoryPath, entryName string) (entry *filer_pb.Entry, err error) {
+	fullPath := util.NewFullPath(parentDirectoryPath, entryName)
+	return filer_pb.GetEntry(s3a, fullPath)
+}
+
 func loadS3config(iam *IdentityAccessManagement, option *S3ApiServerOption) error {
 	return pb.WithCachedGrpcClient(func(grpcConnection *grpc.ClientConn) error {
 		client := filer_pb.NewSeaweedFilerClient(grpcConnection)

weed/s3api/http/header.go

@@ -28,3 +28,9 @@ const (
 	AmzObjectTagging = "X-Amz-Tagging"
 	AmzTagCount      = "x-amz-tagging-count"
 )
+
+// Non-Standard S3 HTTP request constants
+const (
+	AmzIdentityId = "x-amz-identity-id"
+	AmzIsAdmin    = "x-amz-is-admin" // only set to http request header as a context
+)

weed/s3api/s3api_bucket_handlers.go

@@ -4,11 +4,13 @@ import (
 	"context"
 	"encoding/xml"
 	"fmt"
-	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
 	"math"
 	"net/http"
 	"time"
 
+	xhttp "github.com/chrislusf/seaweedfs/weed/s3api/http"
+	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
+
 	"github.com/aws/aws-sdk-go/aws"
 	"github.com/aws/aws-sdk-go/service/s3"
 
@@ -33,9 +35,14 @@ func (s3a *S3ApiServer) ListBucketsHandler(w http.ResponseWriter, r *http.Reques
 		return
 	}
 
+	identityId := r.Header.Get(xhttp.AmzIdentityId)
+
 	var buckets []*s3.Bucket
 	for _, entry := range entries {
 		if entry.IsDirectory {
+			if !s3a.hasAccess(r, entry) {
+				continue
+			}
 			buckets = append(buckets, &s3.Bucket{
 				Name:         aws.String(entry.Name),
 				CreationDate: aws.Time(time.Unix(entry.Attributes.Crtime, 0).UTC()),
@@ -45,8 +52,8 @@ func (s3a *S3ApiServer) ListBucketsHandler(w http.ResponseWriter, r *http.Reques
 
 	response = ListAllMyBucketsResult{
 		Owner: &s3.Owner{
-			ID:          aws.String(""),
+			ID:          aws.String(identityId),
-			DisplayName: aws.String(""),
+			DisplayName: aws.String(identityId),
 		},
 		Buckets: buckets,
 	}
@@ -80,13 +87,25 @@ func (s3a *S3ApiServer) PutBucketHandler(w http.ResponseWriter, r *http.Request)
 		writeErrorResponse(w, s3err.ErrInternalError, r.URL)
 		return
 	}
+	if exist, err := s3a.exists(s3a.option.BucketsPath, bucket, true); err == nil && exist {
+		errCode = s3err.ErrBucketAlreadyExists
+	}
 	if errCode != s3err.ErrNone {
 		writeErrorResponse(w, errCode, r.URL)
 		return
 	}
 
+	fn := func(entry *filer_pb.Entry) {
+		if identityId := r.Header.Get(xhttp.AmzIdentityId); identityId != "" {
+			if entry.Extended == nil {
+				entry.Extended = make(map[string][]byte)
+			}
+			entry.Extended[xhttp.AmzIdentityId] = []byte(identityId)
+		}
+	}
+
 	// create the folder for bucket, but lazily create actual collection
-	if err := s3a.mkdir(s3a.option.BucketsPath, bucket, nil); err != nil {
+	if err := s3a.mkdir(s3a.option.BucketsPath, bucket, fn); err != nil {
 		glog.Errorf("PutBucketHandler mkdir: %v", err)
 		writeErrorResponse(w, s3err.ErrInternalError, r.URL)
 		return
@@ -99,6 +118,11 @@ func (s3a *S3ApiServer) DeleteBucketHandler(w http.ResponseWriter, r *http.Reque
 
 	bucket, _ := getBucketAndObject(r)
 
+	if err := s3a.checkBucket(r, bucket); err != s3err.ErrNone {
+		writeErrorResponse(w, err, r.URL)
+		return
+	}
+
 	err := s3a.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
 
 		// delete collection
@@ -128,28 +152,40 @@ func (s3a *S3ApiServer) HeadBucketHandler(w http.ResponseWriter, r *http.Request
 
 	bucket, _ := getBucketAndObject(r)
 
-	err := s3a.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
+	if err := s3a.checkBucket(r, bucket); err != s3err.ErrNone {
-
+		writeErrorResponse(w, err, r.URL)
-		request := &filer_pb.LookupDirectoryEntryRequest{
-			Directory: s3a.option.BucketsPath,
-			Name:      bucket,
-		}
-
-		glog.V(1).Infof("lookup bucket: %v", request)
-		if _, err := filer_pb.LookupEntry(client, request); err != nil {
-			if err == filer_pb.ErrNotFound {
-				return filer_pb.ErrNotFound
-			}
-			return fmt.Errorf("lookup bucket %s/%s: %v", s3a.option.BucketsPath, bucket, err)
-		}
-
-		return nil
-	})
-
-	if err != nil {
-		writeErrorResponse(w, s3err.ErrNoSuchBucket, r.URL)
 		return
 	}
 
 	writeSuccessResponseEmpty(w)
 }
+
+func (s3a *S3ApiServer) checkBucket(r *http.Request, bucket string) s3err.ErrorCode {
+	entry, err := s3a.getEntry(s3a.option.BucketsPath, bucket)
+	if entry == nil || err == filer_pb.ErrNotFound {
+		return s3err.ErrNoSuchBucket
+	}
+
+	if !s3a.hasAccess(r, entry) {
+		return s3err.ErrAccessDenied
+	}
+	return s3err.ErrNone
+}
+
+func (s3a *S3ApiServer) hasAccess(r *http.Request, entry *filer_pb.Entry) bool {
+	isAdmin := r.Header.Get(xhttp.AmzIsAdmin) != ""
+	if isAdmin {
+		return true
+	}
+	if entry.Extended == nil {
+		return true
+	}
+
+	identityId := r.Header.Get(xhttp.AmzIdentityId)
+	if id, ok := entry.Extended[xhttp.AmzIdentityId]; ok {
+		if identityId != string(id) {
+			return false
+		}
+	}
+	return true
+}

weed/s3api/s3api_object_copy_handlers.go

@@ -2,6 +2,7 @@ package s3api
 
 import (
 	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
 	"net/http"
 	"net/url"
@@ -47,6 +48,7 @@ func (s3a *S3ApiServer) CopyObjectHandler(w http.ResponseWriter, r *http.Request
 	}
 	defer util.CloseResponse(resp)
 
+	glog.V(2).Infof("copy from %s to %s", srcUrl, dstUrl)
 	etag, errCode := s3a.putToFiler(r, dstUrl, resp.Body)
 
 	if errCode != s3err.ErrNone {
@@ -127,6 +129,7 @@ func (s3a *S3ApiServer) CopyObjectPartHandler(w http.ResponseWriter, r *http.Req
 	}
 	defer dataReader.Close()
 
+	glog.V(2).Infof("copy from %s to %s", srcUrl, dstUrl)
 	etag, errCode := s3a.putToFiler(r, dstUrl, dataReader)
 
 	if errCode != s3err.ErrNone {

weed/s3api/s3api_object_handlers.go

@@ -113,12 +113,6 @@ func (s3a *S3ApiServer) DeleteObjectHandler(w http.ResponseWriter, r *http.Reque
 
 	bucket, object := getBucketAndObject(r)
 
-	response, _ := s3a.listFilerEntries(bucket, object, 1, "", "/")
-	if len(response.Contents) != 0 && strings.HasSuffix(object, "/") {
-		w.WriteHeader(http.StatusNoContent)
-		return
-	}
-
 	destUrl := fmt.Sprintf("http://%s%s/%s%s?recursive=true",
 		s3a.option.Filer, s3a.option.BucketsPath, bucket, object)
 
@@ -266,11 +260,6 @@ func (s3a *S3ApiServer) proxyToFiler(w http.ResponseWriter, r *http.Request, des
 
 	resp, postErr := client.Do(proxyReq)
 
-	if resp.ContentLength == -1 && !strings.HasSuffix(destUrl, "/") {
-		writeErrorResponse(w, s3err.ErrNoSuchKey, r.URL)
-		return
-	}
-
 	if postErr != nil {
 		glog.Errorf("post to filer: %v", postErr)
 		writeErrorResponse(w, s3err.ErrInternalError, r.URL)
@@ -278,6 +267,11 @@ func (s3a *S3ApiServer) proxyToFiler(w http.ResponseWriter, r *http.Request, des
 	}
 	defer util.CloseResponse(resp)
 
+	if (resp.ContentLength == -1 || resp.StatusCode == 404) && !strings.HasSuffix(destUrl, "/") {
+		writeErrorResponse(w, s3err.ErrNoSuchKey, r.URL)
+		return
+	}
+
 	responseFn(resp, w)
 
 }
@@ -323,7 +317,7 @@ func (s3a *S3ApiServer) putToFiler(r *http.Request, uploadUrl string, dataReader
 
 	resp_body, ra_err := ioutil.ReadAll(resp.Body)
 	if ra_err != nil {
-		glog.Errorf("upload to filer response read: %v", ra_err)
+		glog.Errorf("upload to filer response read %d: %v", resp.StatusCode, ra_err)
 		return etag, s3err.ErrInternalError
 	}
 	var ret weed_server.FilerPostResult

weed/s3api/s3api_object_multipart_handlers.go

@@ -2,6 +2,7 @@ package s3api
 
 import (
 	"fmt"
+	"github.com/chrislusf/seaweedfs/weed/glog"
 	"github.com/chrislusf/seaweedfs/weed/s3api/s3err"
 	"net/http"
 	"net/url"
@@ -28,13 +29,13 @@ func (s3a *S3ApiServer) NewMultipartUploadHandler(w http.ResponseWriter, r *http
 		Key:    objectKey(aws.String(object)),
 	})
 
+	glog.V(2).Info("NewMultipartUploadHandler", string(encodeResponse(response)), errCode)
+
 	if errCode != s3err.ErrNone {
 		writeErrorResponse(w, errCode, r.URL)
 		return
 	}
 
-	// println("NewMultipartUploadHandler", string(encodeResponse(response)))
-
 	writeSuccessResponseXML(w, encodeResponse(response))
 
 }
@@ -52,7 +53,7 @@ func (s3a *S3ApiServer) CompleteMultipartUploadHandler(w http.ResponseWriter, r
 		UploadId: aws.String(uploadID),
 	})
 
-	// println("CompleteMultipartUploadHandler", string(encodeResponse(response)), errCode)
+	glog.V(2).Info("CompleteMultipartUploadHandler", string(encodeResponse(response)), errCode)
 
 	if errCode != s3err.ErrNone {
 		writeErrorResponse(w, errCode, r.URL)
@@ -81,7 +82,7 @@ func (s3a *S3ApiServer) AbortMultipartUploadHandler(w http.ResponseWriter, r *ht
 		return
 	}
 
-	// println("AbortMultipartUploadHandler", string(encodeResponse(response)))
+	glog.V(2).Info("AbortMultipartUploadHandler", string(encodeResponse(response)))
 
 	writeSuccessResponseXML(w, encodeResponse(response))
 
@@ -114,13 +115,14 @@ func (s3a *S3ApiServer) ListMultipartUploadsHandler(w http.ResponseWriter, r *ht
 		UploadIdMarker: aws.String(uploadIDMarker),
 	})
 
+	glog.V(2).Info("ListMultipartUploadsHandler", string(encodeResponse(response)), errCode)
+
 	if errCode != s3err.ErrNone {
 		writeErrorResponse(w, errCode, r.URL)
 		return
 	}
 
 	// TODO handle encodingType
-	// println("ListMultipartUploadsHandler", string(encodeResponse(response)))
 
 	writeSuccessResponseXML(w, encodeResponse(response))
 }
@@ -147,13 +149,13 @@ func (s3a *S3ApiServer) ListObjectPartsHandler(w http.ResponseWriter, r *http.Re
 		UploadId:         aws.String(uploadID),
 	})
 
+	glog.V(2).Info("ListObjectPartsHandler", string(encodeResponse(response)), errCode)
+
 	if errCode != s3err.ErrNone {
 		writeErrorResponse(w, errCode, r.URL)
 		return
 	}
 
-	// println("ListObjectPartsHandler", string(encodeResponse(response)))
-
 	writeSuccessResponseXML(w, encodeResponse(response))
 
 }

weed/server/common.go

@@ -38,11 +38,13 @@ func init() {
 
 func writeJson(w http.ResponseWriter, r *http.Request, httpStatus int, obj interface{}) (err error) {
 	var bytes []byte
+	if obj != nil {
 		if r.FormValue("pretty") != "" {
 			bytes, err = json.MarshalIndent(obj, "", "  ")
 		} else {
 			bytes, err = json.Marshal(obj)
 		}
+	}
 	if err != nil {
 		return
 	}
@@ -125,6 +127,7 @@ func submitForClientHandler(w http.ResponseWriter, r *http.Request, masterUrl st
 	ar := &operation.VolumeAssignRequest{
 		Count:       count,
 		DataCenter:  r.FormValue("dataCenter"),
+		Rack:        r.FormValue("rack"),
 		Replication: r.FormValue("replication"),
 		Collection:  r.FormValue("collection"),
 		Ttl:         r.FormValue("ttl"),

weed/server/filer_grpc_server.go

@@ -156,7 +156,7 @@ func (fs *FilerServer) CreateEntry(ctx context.Context, req *filer_pb.CreateEntr
 
 	resp = &filer_pb.CreateEntryResponse{}
 
-	chunks, garbage, err2 := fs.cleanupChunks(nil, req.Entry)
+	chunks, garbage, err2 := fs.cleanupChunks(util.Join(req.Directory, req.Entry.Name), nil, req.Entry)
 	if err2 != nil {
 		return &filer_pb.CreateEntryResponse{}, fmt.Errorf("CreateEntry cleanupChunks %s %s: %v", req.Directory, req.Entry.Name, err2)
 	}
@@ -190,7 +190,7 @@ func (fs *FilerServer) UpdateEntry(ctx context.Context, req *filer_pb.UpdateEntr
 		return &filer_pb.UpdateEntryResponse{}, fmt.Errorf("not found %s: %v", fullpath, err)
 	}
 
-	chunks, garbage, err2 := fs.cleanupChunks(entry, req.Entry)
+	chunks, garbage, err2 := fs.cleanupChunks(fullpath, entry, req.Entry)
 	if err2 != nil {
 		return &filer_pb.UpdateEntryResponse{}, fmt.Errorf("UpdateEntry cleanupChunks %s: %v", fullpath, err2)
 	}
@@ -240,7 +240,7 @@ func (fs *FilerServer) UpdateEntry(ctx context.Context, req *filer_pb.UpdateEntr
 	return &filer_pb.UpdateEntryResponse{}, err
 }
 
-func (fs *FilerServer) cleanupChunks(existingEntry *filer.Entry, newEntry *filer_pb.Entry) (chunks, garbage []*filer_pb.FileChunk, err error) {
+func (fs *FilerServer) cleanupChunks(fullpath string, existingEntry *filer.Entry, newEntry *filer_pb.Entry) (chunks, garbage []*filer_pb.FileChunk, err error) {
 
 	// remove old chunks if not included in the new ones
 	if existingEntry != nil {
@@ -257,7 +257,14 @@ func (fs *FilerServer) cleanupChunks(existingEntry *filer.Entry, newEntry *filer
 	garbage = append(garbage, coveredChunks...)
 
 	if newEntry.Attributes != nil {
-		chunks, err = filer.MaybeManifestize(fs.saveAsChunk(newEntry.Attributes.Replication, newEntry.Attributes.Collection, "", "", needle.SecondsToTTL(newEntry.Attributes.TtlSec), false), chunks)
+		so := fs.detectStorageOption(fullpath,
+			newEntry.Attributes.Collection,
+			newEntry.Attributes.Replication,
+			newEntry.Attributes.TtlSec,
+			"",
+			"",
+		)
+		chunks, err = filer.MaybeManifestize(fs.saveAsChunk(so), chunks)
 		if err != nil {
 			// not good, but should be ok
 			glog.V(0).Infof("MaybeManifestize: %v", err)
@@ -275,7 +282,7 @@ func (fs *FilerServer) AppendToEntry(ctx context.Context, req *filer_pb.AppendTo
 
 	fullpath := util.NewFullPath(req.Directory, req.EntryName)
 	var offset int64 = 0
-	entry, err := fs.filer.FindEntry(ctx, util.FullPath(fullpath))
+	entry, err := fs.filer.FindEntry(ctx, fullpath)
 	if err == filer_pb.ErrNotFound {
 		entry = &filer.Entry{
 			FullPath: fullpath,
@@ -297,8 +304,8 @@ func (fs *FilerServer) AppendToEntry(ctx context.Context, req *filer_pb.AppendTo
 	}
 
 	entry.Chunks = append(entry.Chunks, req.Chunks...)
-
+	so := fs.detectStorageOption(string(fullpath), entry.Collection, entry.Replication, entry.TtlSec, "", "")
-	entry.Chunks, err = filer.MaybeManifestize(fs.saveAsChunk(entry.Replication, entry.Collection, "", "", needle.SecondsToTTL(entry.TtlSec), false), entry.Chunks)
+	entry.Chunks, err = filer.MaybeManifestize(fs.saveAsChunk(so), entry.Chunks)
 	if err != nil {
 		// not good, but should be ok
 		glog.V(0).Infof("MaybeManifestize: %v", err)
@@ -323,41 +330,10 @@ func (fs *FilerServer) DeleteEntry(ctx context.Context, req *filer_pb.DeleteEntr
 
 func (fs *FilerServer) AssignVolume(ctx context.Context, req *filer_pb.AssignVolumeRequest) (resp *filer_pb.AssignVolumeResponse, err error) {
 
-	ttlStr := ""
+	so := fs.detectStorageOption(req.Path, req.Collection, req.Replication, req.TtlSec, req.DataCenter, req.Rack)
-	if req.TtlSec > 0 {
-		ttlStr = strconv.Itoa(int(req.TtlSec))
-	}
-	collection, replication, _ := fs.detectCollection(req.Path, req.Collection, req.Replication)
 
-	var altRequest *operation.VolumeAssignRequest
+	assignRequest, altRequest := so.ToAssignRequests(int(req.Count))
 
-	dataCenter := req.DataCenter
-	if dataCenter == "" {
-		dataCenter = fs.option.DataCenter
-	}
-	rack := req.Rack
-	if rack == "" {
-		rack = fs.option.Rack
-	}
-
-	assignRequest := &operation.VolumeAssignRequest{
-		Count:       uint64(req.Count),
-		Replication: replication,
-		Collection:  collection,
-		Ttl:         ttlStr,
-		DataCenter:  dataCenter,
-		Rack:        rack,
-	}
-	if dataCenter != "" || rack != "" {
-		altRequest = &operation.VolumeAssignRequest{
-			Count:       uint64(req.Count),
-			Replication: replication,
-			Collection:  collection,
-			Ttl:         ttlStr,
-			DataCenter:  "",
-			Rack:        "",
-		}
-	}
 	assignResult, err := operation.Assign(fs.filer.GetMaster(), fs.grpcDialOption, assignRequest, altRequest)
 	if err != nil {
 		glog.V(3).Infof("AssignVolume: %v", err)
@@ -374,8 +350,8 @@ func (fs *FilerServer) AssignVolume(ctx context.Context, req *filer_pb.AssignVol
 		Url:         assignResult.Url,
 		PublicUrl:   assignResult.PublicUrl,
 		Auth:        string(assignResult.Auth),
-		Collection:  collection,
+		Collection:  so.Collection,
-		Replication: replication,
+		Replication: so.Replication,
 	}, nil
 }
 

weed/server/filer_server.go

@@ -89,7 +89,7 @@ func NewFilerServer(defaultMux, readonlyMux *http.ServeMux, option *FilerOption)
 		glog.Fatal("master list is required!")
 	}
 
-	fs.filer = filer.NewFiler(option.Masters, fs.grpcDialOption, option.Host, option.Port, option.Collection, option.DefaultReplication, func() {
+	fs.filer = filer.NewFiler(option.Masters, fs.grpcDialOption, option.Host, option.Port, option.Collection, option.DefaultReplication, option.DataCenter, func() {
 		fs.listenersCond.Broadcast()
 	})
 	fs.filer.Cipher = option.Cipher
@@ -114,6 +114,8 @@ func NewFilerServer(defaultMux, readonlyMux *http.ServeMux, option *FilerOption)
 	fs.option.recursiveDelete = v.GetBool("filer.options.recursive_delete")
 	v.SetDefault("filer.options.buckets_folder", "/buckets")
 	fs.filer.DirBucketsPath = v.GetString("filer.options.buckets_folder")
+	// TODO deprecated, will be be removed after 2020-12-31
+	// replaced by https://github.com/chrislusf/seaweedfs/wiki/Path-Specific-Configuration
 	fs.filer.FsyncBuckets = v.GetStringSlice("filer.options.buckets_fsync")
 	fs.filer.LoadConfiguration(v)
 
@@ -131,6 +133,8 @@ func NewFilerServer(defaultMux, readonlyMux *http.ServeMux, option *FilerOption)
 
 	fs.filer.LoadBuckets()
 
+	fs.filer.LoadFilerConf()
+
 	grace.OnInterrupt(func() {
 		fs.filer.Shutdown()
 	})

weed/server/filer_server_handlers.go

@@ -26,11 +26,19 @@ func (fs *FilerServer) filerHandler(w http.ResponseWriter, r *http.Request) {
 		stats.FilerRequestHistogram.WithLabelValues("head").Observe(time.Since(start).Seconds())
 	case "DELETE":
 		stats.FilerRequestCounter.WithLabelValues("delete").Inc()
+		if _, ok := r.URL.Query()["tagging"]; ok {
+			fs.DeleteTaggingHandler(w, r)
+		} else {
 			fs.DeleteHandler(w, r)
+		}
 		stats.FilerRequestHistogram.WithLabelValues("delete").Observe(time.Since(start).Seconds())
 	case "PUT":
 		stats.FilerRequestCounter.WithLabelValues("put").Inc()
+		if _, ok := r.URL.Query()["tagging"]; ok {
+			fs.PutTaggingHandler(w, r)
+		} else {
 			fs.PostHandler(w, r)
+		}
 		stats.FilerRequestHistogram.WithLabelValues("put").Observe(time.Since(start).Seconds())
 	case "POST":
 		stats.FilerRequestCounter.WithLabelValues("post").Inc()

weed/server/filer_server_handlers_read.go

@@ -94,10 +94,15 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request,
 		}
 	}
 
+	// print out the header from extended properties
+	for k, v := range entry.Extended {
+		w.Header().Set(k, string(v))
+	}
+
 	//set tag count
 	if r.Method == "GET" {
 		tagCount := 0
-		for k, _ := range entry.Extended {
+		for k := range entry.Extended {
 			if strings.HasPrefix(k, xhttp.AmzObjectTagging+"-") {
 				tagCount++
 			}

weed/server/filer_server_handlers_tagging.go

@@ -0,0 +1,102 @@
+package weed_server
+
+import (
+	"context"
+	"net/http"
+	"strings"
+
+	"github.com/chrislusf/seaweedfs/weed/glog"
+	"github.com/chrislusf/seaweedfs/weed/storage/needle"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+// add or replace one file Seaweed- prefixed attributes
+// curl -X PUT -H "Seaweed-Name1: value1" http://localhost:8888/path/to/a/file?tagging
+func (fs *FilerServer) PutTaggingHandler(w http.ResponseWriter, r *http.Request) {
+
+	ctx := context.Background()
+
+	path := r.URL.Path
+	if strings.HasSuffix(path, "/") {
+		path = path[:len(path)-1]
+	}
+
+	existingEntry, err := fs.filer.FindEntry(ctx, util.FullPath(path))
+	if err != nil {
+		writeJsonError(w, r, http.StatusNotFound, err)
+		return
+	}
+	if existingEntry == nil {
+		writeJsonError(w, r, http.StatusNotFound, err)
+		return
+	}
+
+	if existingEntry.Extended == nil {
+		existingEntry.Extended = make(map[string][]byte)
+	}
+
+	for header, values := range r.Header {
+		if strings.HasPrefix(header, needle.PairNamePrefix) {
+			for _, value := range values {
+				existingEntry.Extended[header] = []byte(value)
+			}
+		}
+	}
+
+	if dbErr := fs.filer.CreateEntry(ctx, existingEntry, false, false, nil); dbErr != nil {
+		glog.V(0).Infof("failing to update %s tagging : %v", path, dbErr)
+		writeJsonError(w, r, http.StatusInternalServerError, err)
+		return
+	}
+
+	writeJsonQuiet(w, r, http.StatusAccepted, nil)
+	return
+}
+
+// remove all Seaweed- prefixed attributes
+// curl -X DELETE http://localhost:8888/path/to/a/file?tagging
+func (fs *FilerServer) DeleteTaggingHandler(w http.ResponseWriter, r *http.Request) {
+
+	ctx := context.Background()
+
+	path := r.URL.Path
+	if strings.HasSuffix(path, "/") {
+		path = path[:len(path)-1]
+	}
+
+	existingEntry, err := fs.filer.FindEntry(ctx, util.FullPath(path))
+	if err != nil {
+		writeJsonError(w, r, http.StatusNotFound, err)
+		return
+	}
+	if existingEntry == nil {
+		writeJsonError(w, r, http.StatusNotFound, err)
+		return
+	}
+
+	if existingEntry.Extended == nil {
+		existingEntry.Extended = make(map[string][]byte)
+	}
+
+	hasDeletion := false
+	for header, _ := range existingEntry.Extended {
+		if strings.HasPrefix(header, needle.PairNamePrefix) {
+			delete(existingEntry.Extended, header)
+			hasDeletion = true
+		}
+	}
+
+	if !hasDeletion {
+		writeJsonQuiet(w, r, http.StatusNotModified, nil)
+		return
+	}
+
+	if dbErr := fs.filer.CreateEntry(ctx, existingEntry, false, false, nil); dbErr != nil {
+		glog.V(0).Infof("failing to delete %s tagging : %v", path, dbErr)
+		writeJsonError(w, r, http.StatusInternalServerError, err)
+		return
+	}
+
+	writeJsonQuiet(w, r, http.StatusAccepted, nil)
+	return
+}

weed/server/filer_server_handlers_write.go

@@ -29,30 +29,13 @@ type FilerPostResult struct {
 	Url   string `json:"url,omitempty"`
 }
 
-func (fs *FilerServer) assignNewFileInfo(replication, collection, dataCenter, rack, ttlString string, fsync bool) (fileId, urlLocation string, auth security.EncodedJwt, err error) {
+func (fs *FilerServer) assignNewFileInfo(so *operation.StorageOption) (fileId, urlLocation string, auth security.EncodedJwt, err error) {
 
 	stats.FilerRequestCounter.WithLabelValues("assign").Inc()
 	start := time.Now()
 	defer func() { stats.FilerRequestHistogram.WithLabelValues("assign").Observe(time.Since(start).Seconds()) }()
 
-	ar := &operation.VolumeAssignRequest{
+	ar, altRequest := so.ToAssignRequests(1)
-		Count:       1,
-		Replication: replication,
-		Collection:  collection,
-		Ttl:         ttlString,
-		DataCenter:  dataCenter,
-	}
-	var altRequest *operation.VolumeAssignRequest
-	if dataCenter != "" || rack != "" {
-		altRequest = &operation.VolumeAssignRequest{
-			Count:       1,
-			Replication: replication,
-			Collection:  collection,
-			Ttl:         ttlString,
-			DataCenter:  "",
-			Rack:        "",
-		}
-	}
 
 	assignResult, ae := operation.Assign(fs.filer.GetMaster(), fs.grpcDialOption, ar, altRequest)
 	if ae != nil {
@@ -62,7 +45,7 @@ func (fs *FilerServer) assignNewFileInfo(replication, collection, dataCenter, ra
 	}
 	fileId = assignResult.Fid
 	urlLocation = "http://" + assignResult.Url + "/" + assignResult.Fid
-	if fsync {
+	if so.Fsync {
 		urlLocation += "?fsync=true"
 	}
 	auth = assignResult.Auth
@@ -74,25 +57,15 @@ func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) {
 	ctx := context.Background()
 
 	query := r.URL.Query()
-	collection, replication, fsync := fs.detectCollection(r.RequestURI, query.Get("collection"), query.Get("replication"))
+	so := fs.detectStorageOption0(r.RequestURI,
-	dataCenter := query.Get("dataCenter")
+		query.Get("collection"),
-	if dataCenter == "" {
+		query.Get("replication"),
-		dataCenter = fs.option.DataCenter
+		query.Get("ttl"),
-	}
+		query.Get("dataCenter"),
-	rack := query.Get("rack")
+		query.Get("rack"),
-	if dataCenter == "" {
+	)
-		rack = fs.option.Rack
-	}
-	ttlString := r.URL.Query().Get("ttl")
 
-	// read ttl in seconds
+	fs.autoChunk(ctx, w, r, so)
-	ttl, err := needle.ReadTTL(ttlString)
-	ttlSeconds := int32(0)
-	if err == nil {
-		ttlSeconds = int32(ttl.Minutes()) * 60
-	}
-
-	fs.autoChunk(ctx, w, r, replication, collection, dataCenter, rack, ttlSeconds, ttlString, fsync)
 
 }
 
@@ -130,21 +103,12 @@ func (fs *FilerServer) DeleteHandler(w http.ResponseWriter, r *http.Request) {
 	w.WriteHeader(http.StatusNoContent)
 }
 
-func (fs *FilerServer) detectCollection(requestURI, qCollection, qReplication string) (collection, replication string, fsync bool) {
+func (fs *FilerServer) detectStorageOption(requestURI, qCollection, qReplication string, ttlSeconds int32, dataCenter, rack string) *operation.StorageOption {
-	// default
+	collection := util.Nvl(qCollection, fs.option.Collection)
-	collection = fs.option.Collection
+	replication := util.Nvl(qReplication, fs.option.DefaultReplication)
-	replication = fs.option.DefaultReplication
-
-	// get default collection settings
-	if qCollection != "" {
-		collection = qCollection
-	}
-	if qReplication != "" {
-		replication = qReplication
-	}
 
 	// required by buckets folder
-	bucketDefaultReplication := ""
+	bucketDefaultReplication, fsync := "", false
 	if strings.HasPrefix(requestURI, fs.filer.DirBucketsPath+"/") {
 		bucketAndObjectKey := requestURI[len(fs.filer.DirBucketsPath)+1:]
 		t := strings.Index(bucketAndObjectKey, "/")
@@ -160,5 +124,33 @@ func (fs *FilerServer) detectCollection(requestURI, qCollection, qReplication st
 		replication = bucketDefaultReplication
 	}
 
-	return
+	rule := fs.filer.FilerConf.MatchStorageRule(requestURI)
+
+	if ttlSeconds == 0 {
+		ttl, err := needle.ReadTTL(rule.GetTtl())
+		if err != nil {
+			glog.Errorf("fail to parse %s ttl setting %s: %v", rule.LocationPrefix, rule.Ttl, err)
+		}
+		ttlSeconds = int32(ttl.Minutes()) * 60
+	}
+
+	return &operation.StorageOption{
+		Replication:       util.Nvl(replication, rule.Replication),
+		Collection:        util.Nvl(collection, rule.Collection),
+		DataCenter:        util.Nvl(dataCenter, fs.option.DataCenter),
+		Rack:              util.Nvl(rack, fs.option.Rack),
+		TtlSeconds:        ttlSeconds,
+		Fsync:             fsync || rule.Fsync,
+		VolumeGrowthCount: rule.VolumeGrowthCount,
+	}
+}
+
+func (fs *FilerServer) detectStorageOption0(requestURI, qCollection, qReplication string, qTtl string, dataCenter, rack string) *operation.StorageOption {
+
+	ttl, err := needle.ReadTTL(qTtl)
+	if err != nil {
+		glog.Errorf("fail to parse ttl %s: %v", qTtl, err)
+	}
+
+	return fs.detectStorageOption(requestURI, qCollection, qReplication, int32(ttl.Minutes())*60, dataCenter, rack)
 }

weed/server/filer_server_handlers_write_autochunk.go

@@ -21,10 +21,11 @@ import (
 	xhttp "github.com/chrislusf/seaweedfs/weed/s3api/http"
 	"github.com/chrislusf/seaweedfs/weed/security"
 	"github.com/chrislusf/seaweedfs/weed/stats"
+	"github.com/chrislusf/seaweedfs/weed/storage/needle"
 	"github.com/chrislusf/seaweedfs/weed/util"
 )
 
-func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, replication string, collection string, dataCenter string, rack string, ttlSec int32, ttlString string, fsync bool) {
+func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, so *operation.StorageOption) {
 
 	// autoChunking can be set at the command-line level or as a query param. Query param overrides command-line
 	query := r.URL.Query()
@@ -50,10 +51,10 @@ func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *
 		if r.Header.Get("Content-Type") == "" && strings.HasSuffix(r.URL.Path, "/") {
 			reply, err = fs.mkdir(ctx, w, r)
 		} else {
-			reply, md5bytes, err = fs.doPostAutoChunk(ctx, w, r, chunkSize, replication, collection, dataCenter, rack, ttlSec, ttlString, fsync)
+			reply, md5bytes, err = fs.doPostAutoChunk(ctx, w, r, chunkSize, so)
 		}
 	} else {
-		reply, md5bytes, err = fs.doPutAutoChunk(ctx, w, r, chunkSize, replication, collection, dataCenter, rack, ttlSec, ttlString, fsync)
+		reply, md5bytes, err = fs.doPutAutoChunk(ctx, w, r, chunkSize, so)
 	}
 	if err != nil {
 		writeJsonError(w, r, http.StatusInternalServerError, err)
@@ -65,7 +66,7 @@ func (fs *FilerServer) autoChunk(ctx context.Context, w http.ResponseWriter, r *
 	}
 }
 
-func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, replication string, collection string, dataCenter string, rack string, ttlSec int32, ttlString string, fsync bool) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
+func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
 
 	multipartReader, multipartReaderErr := r.MultipartReader()
 	if multipartReaderErr != nil {
@@ -86,46 +87,46 @@ func (fs *FilerServer) doPostAutoChunk(ctx context.Context, w http.ResponseWrite
 		contentType = ""
 	}
 
-	fileChunks, md5Hash, chunkOffset, err := fs.uploadReaderToChunks(w, r, part1, chunkSize, replication, collection, dataCenter, rack, ttlString, fileName, contentType, fsync)
+	fileChunks, md5Hash, chunkOffset, err := fs.uploadReaderToChunks(w, r, part1, chunkSize, fileName, contentType, so)
 	if err != nil {
 		return nil, nil, err
 	}
 
-	fileChunks, replyerr = filer.MaybeManifestize(fs.saveAsChunk(replication, collection, dataCenter, rack, ttlString, fsync), fileChunks)
+	fileChunks, replyerr = filer.MaybeManifestize(fs.saveAsChunk(so), fileChunks)
 	if replyerr != nil {
 		glog.V(0).Infof("manifestize %s: %v", r.RequestURI, replyerr)
 		return
 	}
 
 	md5bytes = md5Hash.Sum(nil)
-	filerResult, replyerr = fs.saveMetaData(ctx, r, fileName, replication, collection, ttlSec, contentType, md5bytes, fileChunks, chunkOffset)
+	filerResult, replyerr = fs.saveMetaData(ctx, r, fileName, contentType, so, md5bytes, fileChunks, chunkOffset)
 
 	return
 }
 
-func (fs *FilerServer) doPutAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, replication string, collection string, dataCenter string, rack string, ttlSec int32, ttlString string, fsync bool) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
+func (fs *FilerServer) doPutAutoChunk(ctx context.Context, w http.ResponseWriter, r *http.Request, chunkSize int32, so *operation.StorageOption) (filerResult *FilerPostResult, md5bytes []byte, replyerr error) {
 
 	fileName := ""
 	contentType := ""
 
-	fileChunks, md5Hash, chunkOffset, err := fs.uploadReaderToChunks(w, r, r.Body, chunkSize, replication, collection, dataCenter, rack, ttlString, fileName, contentType, fsync)
+	fileChunks, md5Hash, chunkOffset, err := fs.uploadReaderToChunks(w, r, r.Body, chunkSize, fileName, contentType, so)
 	if err != nil {
 		return nil, nil, err
 	}
 
-	fileChunks, replyerr = filer.MaybeManifestize(fs.saveAsChunk(replication, collection, dataCenter, rack, ttlString, fsync), fileChunks)
+	fileChunks, replyerr = filer.MaybeManifestize(fs.saveAsChunk(so), fileChunks)
 	if replyerr != nil {
 		glog.V(0).Infof("manifestize %s: %v", r.RequestURI, replyerr)
 		return
 	}
 
 	md5bytes = md5Hash.Sum(nil)
-	filerResult, replyerr = fs.saveMetaData(ctx, r, fileName, replication, collection, ttlSec, contentType, md5bytes, fileChunks, chunkOffset)
+	filerResult, replyerr = fs.saveMetaData(ctx, r, fileName, contentType, so, md5bytes, fileChunks, chunkOffset)
 
 	return
 }
 
-func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileName string, replication string, collection string, ttlSec int32, contentType string, md5bytes []byte, fileChunks []*filer_pb.FileChunk, chunkOffset int64) (filerResult *FilerPostResult, replyerr error) {
+func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileName string, contentType string, so *operation.StorageOption, md5bytes []byte, fileChunks []*filer_pb.FileChunk, chunkOffset int64) (filerResult *FilerPostResult, replyerr error) {
 
 	// detect file mode
 	modeStr := r.URL.Query().Get("mode")
@@ -162,9 +163,9 @@ func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileNa
 			Mode:        os.FileMode(mode),
 			Uid:         OS_UID,
 			Gid:         OS_GID,
-			Replication: replication,
+			Replication: so.Replication,
-			Collection:  collection,
+			Collection:  so.Collection,
-			TtlSec:      ttlSec,
+			TtlSec:      so.TtlSeconds,
 			Mime:        contentType,
 			Md5:         md5bytes,
 			FileSize:    uint64(chunkOffset),
@@ -177,8 +178,18 @@ func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileNa
 		Size: chunkOffset,
 	}
 
+	if entry.Extended == nil {
+		entry.Extended = make(map[string][]byte)
+	}
+
 	fs.saveAmzMetaData(r, entry)
 
+	for k, v := range r.Header {
+		if len(v) > 0 && strings.HasPrefix(k, needle.PairNamePrefix) {
+			entry.Extended[k] = []byte(v[0])
+		}
+	}
+
 	if dbErr := fs.filer.CreateEntry(ctx, entry, false, false, nil); dbErr != nil {
 		fs.filer.DeleteChunks(entry.Chunks)
 		replyerr = dbErr
@@ -188,7 +199,7 @@ func (fs *FilerServer) saveMetaData(ctx context.Context, r *http.Request, fileNa
 	return filerResult, replyerr
 }
 
-func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Request, reader io.Reader, chunkSize int32, replication string, collection string, dataCenter string, rack string, ttlString string, fileName string, contentType string, fsync bool) ([]*filer_pb.FileChunk, hash.Hash, int64, error) {
+func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Request, reader io.Reader, chunkSize int32, fileName, contentType string, so *operation.StorageOption) ([]*filer_pb.FileChunk, hash.Hash, int64, error) {
 	var fileChunks []*filer_pb.FileChunk
 
 	md5Hash := md5.New()
@@ -200,7 +211,7 @@ func (fs *FilerServer) uploadReaderToChunks(w http.ResponseWriter, r *http.Reque
 		limitedReader := io.LimitReader(partReader, int64(chunkSize))
 
 		// assign one file id for one chunk
-		fileId, urlLocation, auth, assignErr := fs.assignNewFileInfo(replication, collection, dataCenter, rack, ttlString, fsync)
+		fileId, urlLocation, auth, assignErr := fs.assignNewFileInfo(so)
 		if assignErr != nil {
 			return nil, nil, 0, assignErr
 		}
@@ -244,11 +255,11 @@ func (fs *FilerServer) doUpload(urlLocation string, w http.ResponseWriter, r *ht
 	return uploadResult, err
 }
 
-func (fs *FilerServer) saveAsChunk(replication string, collection string, dataCenter string, rack string, ttlString string, fsync bool) filer.SaveDataAsChunkFunctionType {
+func (fs *FilerServer) saveAsChunk(so *operation.StorageOption) filer.SaveDataAsChunkFunctionType {
 
 	return func(reader io.Reader, name string, offset int64) (*filer_pb.FileChunk, string, string, error) {
 		// assign one file id for one chunk
-		fileId, urlLocation, auth, assignErr := fs.assignNewFileInfo(replication, collection, dataCenter, rack, ttlString, fsync)
+		fileId, urlLocation, auth, assignErr := fs.assignNewFileInfo(so)
 		if assignErr != nil {
 			return nil, "", "", assignErr
 		}
@@ -259,7 +270,7 @@ func (fs *FilerServer) saveAsChunk(replication string, collection string, dataCe
 			return nil, "", "", uploadErr
 		}
 
-		return uploadResult.ToPbFileChunk(fileId, offset), collection, replication, nil
+		return uploadResult.ToPbFileChunk(fileId, offset), so.Collection, so.Replication, nil
 	}
 }
 
@@ -314,10 +325,6 @@ func (fs *FilerServer) mkdir(ctx context.Context, w http.ResponseWriter, r *http
 
 func (fs *FilerServer) saveAmzMetaData(r *http.Request, entry *filer.Entry) {
 
-	if entry.Extended == nil {
-		entry.Extended = make(map[string][]byte)
-	}
-
 	if sc := r.Header.Get(xhttp.AmzStorageClass); sc != "" {
 		entry.Extended[xhttp.AmzStorageClass] = []byte(sc)
 	}

weed/server/filer_server_handlers_write_cipher.go

@@ -16,12 +16,12 @@ import (
 )
 
 // handling single chunk POST or PUT upload
-func (fs *FilerServer) encrypt(ctx context.Context, w http.ResponseWriter, r *http.Request, replication string, collection string, dataCenter string, rack string, ttlSeconds int32, ttlString string, fsync bool) (filerResult *FilerPostResult, err error) {
+func (fs *FilerServer) encrypt(ctx context.Context, w http.ResponseWriter, r *http.Request, so *operation.StorageOption) (filerResult *FilerPostResult, err error) {
 
-	fileId, urlLocation, auth, err := fs.assignNewFileInfo(replication, collection, dataCenter, rack, ttlString, fsync)
+	fileId, urlLocation, auth, err := fs.assignNewFileInfo(so)
 
 	if err != nil || fileId == "" || urlLocation == "" {
-		return nil, fmt.Errorf("fail to allocate volume for %s, collection:%s, datacenter:%s", r.URL.Path, collection, dataCenter)
+		return nil, fmt.Errorf("fail to allocate volume for %s, collection:%s, datacenter:%s", r.URL.Path, so.Collection, so.DataCenter)
 	}
 
 	glog.V(4).Infof("write %s to %v", r.URL.Path, urlLocation)
@@ -65,9 +65,9 @@ func (fs *FilerServer) encrypt(ctx context.Context, w http.ResponseWriter, r *ht
 			Mode:        0660,
 			Uid:         OS_UID,
 			Gid:         OS_GID,
-			Replication: replication,
+			Replication: so.Replication,
-			Collection:  collection,
+			Collection:  so.Collection,
-			TtlSec:      ttlSeconds,
+			TtlSec:      so.TtlSeconds,
 			Mime:        pu.MimeType,
 			Md5:         util.Base64Md5ToBytes(pu.ContentMd5),
 		},

weed/server/master_grpc_server.go

@@ -88,6 +88,7 @@ func (ms *MasterServer) SendHeartbeat(stream master_pb.Seaweed_SendHeartbeatServ
 		message := &master_pb.VolumeLocation{
 			Url:        dn.Url(),
 			PublicUrl:  dn.PublicUrl,
+			DataCenter: string(dn.GetDataCenter().Id()),
 		}
 		if len(heartbeat.NewVolumes) > 0 || len(heartbeat.DeletedVolumes) > 0 {
 			// process delta volume ids if exists for fast volume id updates
@@ -148,7 +149,6 @@ func (ms *MasterServer) SendHeartbeat(stream master_pb.Seaweed_SendHeartbeatServ
 			}
 
 		}
-
 		if len(message.NewVids) > 0 || len(message.DeletedVids) > 0 {
 			ms.clientChansLock.RLock()
 			for host, ch := range ms.clientChans {

weed/server/master_server.go

@@ -93,7 +93,7 @@ func NewMasterServer(r *mux.Router, option *MasterOption, peers []string) *Maste
 		preallocateSize: preallocateSize,
 		clientChans:     make(map[string]chan *master_pb.VolumeLocation),
 		grpcDialOption:  grpcDialOption,
-		MasterClient:    wdclient.NewMasterClient(grpcDialOption, "master", option.Host, 0, peers),
+		MasterClient:    wdclient.NewMasterClient(grpcDialOption, "master", option.Host, 0, "", peers),
 		adminLocks:      NewAdminLocks(),
 	}
 	ms.bounedLeaderChan = make(chan int, 16)

weed/server/volume_grpc_client_to_master.go

@@ -203,6 +203,7 @@ func (vs *VolumeServer) doHeartbeat(masterNode, masterGrpcAddress string, grpcDi
 			}
 		case <-volumeTickChan:
 			glog.V(4).Infof("volume server %s:%d heartbeat", vs.store.Ip, vs.store.Port)
+			vs.store.MaybeAdjustVolumeMax()
 			if err = stream.Send(vs.store.CollectHeartbeat()); err != nil {
 				glog.V(0).Infof("Volume Server Failed to talk with master %s: %v", masterNode, err)
 				return "", err

weed/shell/command_fs_configure.go

@@ -0,0 +1,151 @@
+package shell
+
+import (
+	"bytes"
+	"flag"
+	"fmt"
+	"io"
+	"math"
+	"net/http"
+	"strings"
+
+	"github.com/chrislusf/seaweedfs/weed/filer"
+	"github.com/chrislusf/seaweedfs/weed/pb/filer_pb"
+	"github.com/chrislusf/seaweedfs/weed/storage/super_block"
+	"github.com/chrislusf/seaweedfs/weed/util"
+)
+
+func init() {
+	Commands = append(Commands, &commandFsConfigure{})
+}
+
+type commandFsConfigure struct {
+}
+
+func (c *commandFsConfigure) Name() string {
+	return "fs.configure"
+}
+
+func (c *commandFsConfigure) Help() string {
+	return `configure and apply storage options for each location
+
+	# see the current configuration file content
+	fs.configure
+
+	# trying the changes and see the possible configuration file content
+	fs.configure -locationPrfix=/my/folder -collection=abc
+	fs.configure -locationPrfix=/my/folder -collection=abc -ttl=7d
+
+	# example: configure adding only 1 physical volume for each bucket collection
+	fs.configure -locationPrfix=/buckets/ -volumeGrowthCount=1
+
+	# apply the changes
+	fs.configure -locationPrfix=/my/folder -collection=abc -apply
+
+	# delete the changes
+	fs.configure -locationPrfix=/my/folder -delete -apply
+
+`
+}
+
+func (c *commandFsConfigure) Do(args []string, commandEnv *CommandEnv, writer io.Writer) (err error) {
+
+	fsConfigureCommand := flag.NewFlagSet(c.Name(), flag.ContinueOnError)
+	locationPrefix := fsConfigureCommand.String("locationPrefix", "", "path prefix, required to update the path-specific configuration")
+	collection := fsConfigureCommand.String("collection", "", "assign writes to this collection")
+	replication := fsConfigureCommand.String("replication", "", "assign writes with this replication")
+	ttl := fsConfigureCommand.String("ttl", "", "assign writes with this ttl")
+	fsync := fsConfigureCommand.Bool("fsync", false, "fsync for the writes")
+	volumeGrowthCount := fsConfigureCommand.Int("volumeGrowthCount", 0, "the number of physical volumes to add if no writable volumes")
+	isDelete := fsConfigureCommand.Bool("delete", false, "delete the configuration by locationPrefix")
+	apply := fsConfigureCommand.Bool("apply", false, "update and apply filer configuration")
+	if err = fsConfigureCommand.Parse(args); err != nil {
+		return nil
+	}
+
+	var buf bytes.Buffer
+	if err = commandEnv.WithFilerClient(func(client filer_pb.SeaweedFilerClient) error {
+
+		request := &filer_pb.LookupDirectoryEntryRequest{
+			Directory: filer.DirectoryEtc,
+			Name:      filer.FilerConfName,
+		}
+		respLookupEntry, err := filer_pb.LookupEntry(client, request)
+		if err != nil {
+			return err
+		}
+
+		return filer.StreamContent(commandEnv.MasterClient, &buf, respLookupEntry.Entry.Chunks, 0, math.MaxInt64)
+
+	}); err != nil {
+		return err
+	}
+
+	fc := filer.NewFilerConf()
+	if err = fc.LoadFromBytes(buf.Bytes()); err != nil {
+		return err
+	}
+
+	if *locationPrefix != "" {
+		locConf := &filer_pb.FilerConf_PathConf{
+			LocationPrefix:    *locationPrefix,
+			Collection:        *collection,
+			Replication:       *replication,
+			Ttl:               *ttl,
+			Fsync:             *fsync,
+			VolumeGrowthCount: uint32(*volumeGrowthCount),
+		}
+
+		// check collection
+		if *collection != "" && strings.HasPrefix(*locationPrefix, "/buckets/") {
+			return fmt.Errorf("one s3 bucket goes to one collection and not customizable.")
+		}
+
+		// check replication
+		if *replication != "" {
+			rp, err := super_block.NewReplicaPlacementFromString(*replication)
+			if err != nil {
+				return fmt.Errorf("parse replication %s: %v", *replication, err)
+			}
+			if *volumeGrowthCount % rp.GetCopyCount() != 0 {
+				return fmt.Errorf("volumeGrowthCount %d should be devided by replication copy count %d", *volumeGrowthCount, rp.GetCopyCount())
+			}
+		}
+
+		// save it
+		if *isDelete {
+			fc.DeleteLocationConf(*locationPrefix)
+		} else {
+			fc.AddLocationConf(locConf)
+		}
+	}
+
+	buf.Reset()
+	fc.ToText(&buf)
+
+	fmt.Fprintf(writer, string(buf.Bytes()))
+	fmt.Fprintln(writer)
+
+	if *apply {
+
+		target := fmt.Sprintf("http://%s:%d%s/%s", commandEnv.option.FilerHost, commandEnv.option.FilerPort, filer.DirectoryEtc, filer.FilerConfName)
+
+		// set the HTTP method, url, and request body
+		req, err := http.NewRequest(http.MethodPut, target, &buf)
+		if err != nil {
+			return err
+		}
+
+		// set the request header Content-Type for json
+		req.Header.Set("Content-Type", "text/plain; charset=utf-8")
+		resp, err := http.DefaultClient.Do(req)
+		if err != nil {
+			return err
+		}
+		util.CloseResponse(resp)
+
+	}
+
+	return nil
+
+}

weed/shell/command_volume_fix_replication.go

@@ -369,18 +369,20 @@ func countReplicas(replicas []*VolumeReplica) (diffDc, diffRack, diffNode map[st
 
 func pickOneReplicaToDelete(replicas []*VolumeReplica, replicaPlacement *super_block.ReplicaPlacement) *VolumeReplica {
 
-	allSame := true
+	sort.Slice(replicas, func(i, j int) bool {
-	oldest := replicas[0]
+		a, b := replicas[i], replicas[j]
-	for _, replica := range replicas {
+		if a.info.CompactRevision != b.info.CompactRevision {
-		if replica.info.ModifiedAtSecond < oldest.info.ModifiedAtSecond {
+			return a.info.CompactRevision < b.info.CompactRevision
-			oldest = replica
-			allSame = false
 		}
+		if a.info.ModifiedAtSecond != b.info.ModifiedAtSecond {
+			return a.info.ModifiedAtSecond < b.info.ModifiedAtSecond
 		}
-	if !allSame {
+		if a.info.Size != b.info.Size {
-		return oldest
+			return a.info.Size < b.info.Size
 		}
+		return false
+	})
+
+	return replicas[0]
 
-	// TODO what if all the replicas have the same timestamp?
-	return oldest
 }

weed/shell/commands.go

@@ -45,7 +45,7 @@ var (
 func NewCommandEnv(options ShellOptions) *CommandEnv {
 	ce := &CommandEnv{
 		env:          make(map[string]string),
-		MasterClient: wdclient.NewMasterClient(options.GrpcDialOption, pb.AdminShellClient, "", 0, strings.Split(*options.Masters, ",")),
+		MasterClient: wdclient.NewMasterClient(options.GrpcDialOption, pb.AdminShellClient, "", 0, "", strings.Split(*options.Masters, ",")),
 		option:       options,
 	}
 	ce.locker = exclusive_locks.NewExclusiveLocker(ce.MasterClient)

weed/storage/disk_location.go

@@ -19,6 +19,7 @@ import (
 type DiskLocation struct {
 	Directory              string
 	MaxVolumeCount         int
+	OriginalMaxVolumeCount int
 	MinFreeSpacePercent    float32
 	volumes                map[needle.VolumeId]*Volume
 	volumesLock            sync.RWMutex
@@ -31,7 +32,7 @@ type DiskLocation struct {
 }
 
 func NewDiskLocation(dir string, maxVolumeCount int, minFreeSpacePercent float32) *DiskLocation {
-	location := &DiskLocation{Directory: dir, MaxVolumeCount: maxVolumeCount, MinFreeSpacePercent: minFreeSpacePercent}
+	location := &DiskLocation{Directory: dir, MaxVolumeCount: maxVolumeCount, OriginalMaxVolumeCount: maxVolumeCount, MinFreeSpacePercent: minFreeSpacePercent}
 	location.volumes = make(map[needle.VolumeId]*Volume)
 	location.ecVolumes = make(map[needle.VolumeId]*erasure_coding.EcVolume)
 	go location.CheckDiskSpace()

weed/storage/store.go

@@ -461,7 +461,8 @@ func (s *Store) GetVolumeSizeLimit() uint64 {
 func (s *Store) MaybeAdjustVolumeMax() (hasChanges bool) {
 	volumeSizeLimit := s.GetVolumeSizeLimit()
 	for _, diskLocation := range s.Locations {
-		if diskLocation.MaxVolumeCount == 0 {
+		if diskLocation.OriginalMaxVolumeCount == 0 {
+			currentMaxVolumeCount := diskLocation.MaxVolumeCount
 			diskStatus := stats.NewDiskStatus(diskLocation.Directory)
 			unusedSpace := diskLocation.UnUsedSpace(volumeSizeLimit)
 			unclaimedSpaces := int64(diskStatus.Free) - int64(unusedSpace)
@@ -471,9 +472,9 @@ func (s *Store) MaybeAdjustVolumeMax() (hasChanges bool) {
 				maxVolumeCount += int(uint64(unclaimedSpaces)/volumeSizeLimit) - 1
 			}
 			diskLocation.MaxVolumeCount = maxVolumeCount
-			glog.V(0).Infof("disk %s max %d unclaimedSpace:%dMB, unused:%dMB volumeSizeLimit:%dMB",
+			glog.V(2).Infof("disk %s max %d unclaimedSpace:%dMB, unused:%dMB volumeSizeLimit:%dMB",
 				diskLocation.Directory, maxVolumeCount, unclaimedSpaces/1024/1024, unusedSpace/1024/1024, volumeSizeLimit/1024/1024)
-			hasChanges = true
+			hasChanges = hasChanges || currentMaxVolumeCount != diskLocation.MaxVolumeCount
 		}
 	}
 	return

weed/topology/store_replicate.go

@@ -81,6 +81,7 @@ func ReplicatedWrite(masterNode string, s *storage.Store, volumeId needle.Volume
 			}
 
 			// volume server do not know about encryption
+			// TODO optimize here to compress data only once
 			_, err := operation.UploadData(u.String(), string(n.Name), false, n.Data, n.IsCompressed(), string(n.Mime), pairMap, jwt)
 			return err
 		}); err != nil {

weed/topology/volume_layout.go

@@ -132,6 +132,8 @@ func NewVolumeLayout(rp *super_block.ReplicaPlacement, ttl *needle.TTL, volumeSi
 }
 
 func (vl *VolumeLayout) String() string {
+	vl.accessLock.RLock()
+	defer vl.accessLock.RUnlock()
 	return fmt.Sprintf("rp:%v, ttl:%v, vid2location:%v, writables:%v, volumeSizeLimit:%v", vl.rp, vl.ttl, vl.vid2location, vl.writables, vl.volumeSizeLimit)
 }
 

weed/util/constants.go

@@ -5,7 +5,7 @@ import (
 )
 
 var (
-	VERSION = fmt.Sprintf("%s %d.%02d", sizeLimit, 2, 8)
+	VERSION = fmt.Sprintf("%s %d.%02d", sizeLimit, 2, 11)
 	COMMIT  = ""
 )
 

weed/util/http_util.go

@@ -1,7 +1,6 @@
 package util
 
 import (
-	"bytes"
 	"compress/gzip"
 	"encoding/json"
 	"errors"
@@ -29,22 +28,6 @@ func init() {
 	}
 }
 
-func PostBytes(url string, body []byte) ([]byte, error) {
-	r, err := client.Post(url, "", bytes.NewReader(body))
-	if err != nil {
-		return nil, fmt.Errorf("Post to %s: %v", url, err)
-	}
-	defer r.Body.Close()
-	b, err := ioutil.ReadAll(r.Body)
-	if err != nil {
-		return nil, fmt.Errorf("Read response body: %v", err)
-	}
-	if r.StatusCode >= 400 {
-		return nil, fmt.Errorf("%s: %s", url, r.Status)
-	}
-	return b, nil
-}
-
 func Post(url string, values url.Values) ([]byte, error) {
 	r, err := client.PostForm(url, values)
 	if err != nil {
@@ -370,7 +353,6 @@ func ReadUrlAsReaderCloser(fileUrl string, rangeHeader string) (io.ReadCloser, e
 	if err != nil {
 		return nil, err
 	}
-	defer CloseResponse(r)
 	if r.StatusCode >= 400 {
 		return nil, fmt.Errorf("%s: %s", fileUrl, r.Status)
 	}

weed/util/retry.go

@@ -29,3 +29,13 @@ func Retry(name string, job func() error) (err error) {
 	}
 	return err
 }
+
+// return the first non empty string
+func Nvl(values ...string) string {
+	for _, s := range values {
+		if s != "" {
+			return s
+		}
+	}
+	return ""
+}

weed/wdclient/masterclient.go

@@ -24,14 +24,14 @@ type MasterClient struct {
 	vidMap
 }
 
-func NewMasterClient(grpcDialOption grpc.DialOption, clientType string, clientHost string, clientGrpcPort uint32, masters []string) *MasterClient {
+func NewMasterClient(grpcDialOption grpc.DialOption, clientType string, clientHost string, clientGrpcPort uint32, clientDataCenter string, masters []string) *MasterClient {
 	return &MasterClient{
 		clientType:     clientType,
 		clientHost:     clientHost,
 		grpcPort:       clientGrpcPort,
 		masters:        masters,
 		grpcDialOption: grpcDialOption,
-		vidMap:         newVidMap(),
+		vidMap:         newVidMap(clientDataCenter),
 	}
 }
 
@@ -89,7 +89,7 @@ func (mc *MasterClient) tryAllMasters() {
 		}
 
 		mc.currentMaster = ""
-		mc.vidMap = newVidMap()
+		mc.vidMap = newVidMap("")
 	}
 }
 
@@ -132,6 +132,7 @@ func (mc *MasterClient) tryConnectToMaster(master string) (nextHintedLeader stri
 			loc := Location{
 				Url:        volumeLocation.Url,
 				PublicUrl:  volumeLocation.PublicUrl,
+				DataCenter: volumeLocation.DataCenter,
 			}
 			for _, newVid := range volumeLocation.NewVids {
 				glog.V(1).Infof("%s: %s masterClient adds volume %d", mc.clientType, loc.Url, newVid)

weed/wdclient/vid_map.go

@@ -18,18 +18,20 @@ const (
 type Location struct {
 	Url        string `json:"url,omitempty"`
 	PublicUrl  string `json:"publicUrl,omitempty"`
+	DataCenter string `json:"dataCenter,omitempty"`
 }
 
 type vidMap struct {
 	sync.RWMutex
 	vid2Locations map[uint32][]Location
-
+	DataCenter    string
 	cursor        int32
 }
 
-func newVidMap() vidMap {
+func newVidMap(dataCenter string) vidMap {
 	return vidMap{
 		vid2Locations: make(map[uint32][]Location),
+		DataCenter:    dataCenter,
 		cursor:        -1,
 	}
 }
@@ -56,7 +58,11 @@ func (vc *vidMap) LookupVolumeServerUrl(vid string) (serverUrls []string, err er
 		return nil, fmt.Errorf("volume %d not found", id)
 	}
 	for _, loc := range locations {
+		if vc.DataCenter == "" || loc.DataCenter == "" || vc.DataCenter != loc.DataCenter {
 			serverUrls = append(serverUrls, loc.Url)
+		} else {
+			serverUrls = append([]string{loc.Url}, serverUrls...)
+		}
 	}
 	return
 }

weed/wdclient/vid_map_test.go

@@ -45,7 +45,7 @@ func TestLocationIndex(t *testing.T) {
 	mustOk(7, maxCursorIndex, 0)
 
 	// test with constructor
-	vm = newVidMap()
+	vm = newVidMap("")
 	length := 7
 	for i := 0; i < 100; i++ {
 		got, err := vm.getLocationIndex(length)