We need to create a "Community" section like this: https://coredns.io/community/

[ root@A01-R02-I182-64-CX5PD92 cmd]# sh build. shgo: finding github. com/kylelemons/godebug v1.1.0
go: finding github. com/facebookgo/stack v0.0.0-20160209184415-751773369052
go: finding github. com/smartystreets/goconvey v0.0.0-20190330032615-68dc04aab96a go: finding github. com/parnurzeal/gorequest vo.2.15
go: finding github. com/google/btree v1.0.0
go: finding github. com/jacobsa/fuse v0.0.0-20180417054321-cd3959611bcb go: finding github. com/elazarl/goproxy/ext v0.0.0-20190421051319-9d40249d3c2f go: finding github. com/tecbot/gorocksdb v0.0.0-20190519120508-025c3cf4ffb4
go: finding bazil. org/fuse v0.0.0-20180421153158-65cc252bf669
go: golang. org/x/[email protected]: unrecognized import path "golang. org/x/net"(https fetch: Get https://golang. org/x/net? go-get=1: dial tcp 216.239.37.1:443: connect: connection timed out)
go: finding github. com/tiglabs/raft v0.0.0-20190131082128-45667fcdb8b8
go: finding github. com/elazarl/goproxy v0.0.0-20190421051319-9d40249d3c2f go: finding github. com/prometheus/client_golang v1.0.0
go: finding github. com/facebookgo/subset v0.0.0-20150612182917-8dac2c3c4870
go: finding github. com/moul/http2curl v1.0.0
go: finding github. com/facebookgo/ensure v0.0.0-20160127193407-b4ab57deab51
go: error loading module requirements

There is no description of REST_SIZE in the datanode config json file.
Is this the maximum usable filesystem size for chubaofs? Measured in bytes?

sh build.sh

github.com/chubaofs/chubaofs/vendor/github.com/tecbot/gorocksdb

/usr/bin/ld: cannot find -lbz2
/usr/bin/ld: cannot find -lsnappy
collect2: 错误：ld 返回 1

S3-like object storage interface to ChubaoFS

[root@A01-R02-I182-64-CX5PD92 cmd]# sh build.sh

github.com/chubaofs/chubaofs/vendor/github.com/tecbot/gorocksdb

/usr/local/lib/librocksdb.a(db_impl.o): In function ZSTD_Supported': /export/servers/temp/rocksdb/./util/compression.h:88: undefined reference to ZSTD_versionNumber'
/usr/local/lib/librocksdb.a(options_helper.o): In function ZSTD_Supported': /export/servers/temp/rocksdb/./util/compression.h:88: undefined reference to ZSTD_versionNumber'
/usr/local/lib/librocksdb.a(format.o): In function ZSTD_Uncompress': /export/servers/temp/rocksdb/./util/compression.h:783: undefined reference to ZSTD_createDCtx'
/export/servers/temp/rocksdb/./util/compression.h:786: undefined reference to ZSTD_decompress_usingDict' /export/servers/temp/rocksdb/./util/compression.h:787: undefined reference to ZSTD_freeDCtx'
/usr/local/lib/librocksdb.a(format.o): In function LZ4_Uncompress': /export/servers/temp/rocksdb/./util/compression.h:617: undefined reference to LZ4_createStreamDecode'
/export/servers/temp/rocksdb/./util/compression.h:624: undefined reference to LZ4_decompress_safe_continue' /export/servers/temp/rocksdb/./util/compression.h:625: undefined reference to LZ4_freeStreamDecode'
/export/servers/temp/rocksdb/./util/compression.h:617: undefined reference to LZ4_createStreamDecode' /export/servers/temp/rocksdb/./util/compression.h:624: undefined reference to LZ4_decompress_safe_continue'
/export/servers/temp/rocksdb/./util/compression.h:625: undefined reference to LZ4_freeStreamDecode' /export/servers/temp/rocksdb/./util/compression.h:620: undefined reference to LZ4_setStreamDecode'
/export/servers/temp/rocksdb/./util/compression.h:620: undefined reference to LZ4_setStreamDecode' /usr/local/lib/librocksdb.a(column_family.o): In function ZSTD_Supported':
/export/servers/temp/rocksdb/./util/compression.h:88: undefined reference to ZSTD_versionNumber' /export/servers/temp/rocksdb/./util/compression.h:88: undefined reference to ZSTD_versionNumber'
/usr/local/lib/librocksdb.a(block_based_table_builder.o): In function ZSTD_Compress': /export/servers/temp/rocksdb/./util/compression.h:746: undefined reference to ZSTD_compressBound'
/export/servers/temp/rocksdb/./util/compression.h:750: undefined reference to ZSTD_createCCtx' /export/servers/temp/rocksdb/./util/compression.h:753: undefined reference to ZSTD_compress_usingDict'
/export/servers/temp/rocksdb/./util/compression.h:754: undefined reference to ZSTD_freeCCtx' /usr/local/lib/librocksdb.a(block_based_table_builder.o): In function LZ4_Compress':
/export/servers/temp/rocksdb/./util/compression.h:552: undefined reference to LZ4_compressBound' /export/servers/temp/rocksdb/./util/compression.h:557: undefined reference to LZ4_createStream'
/export/servers/temp/rocksdb/./util/compression.h:565: undefined reference to LZ4_compress_fast_continue' /export/servers/temp/rocksdb/./util/compression.h:571: undefined reference to LZ4_freeStream'
/usr/local/lib/librocksdb.a(block_based_table_builder.o): In function LZ4HC_Compress': /export/servers/temp/rocksdb/./util/compression.h:671: undefined reference to LZ4_compressBound'
/export/servers/temp/rocksdb/./util/compression.h:676: undefined reference to LZ4_createStreamHC' /export/servers/temp/rocksdb/./util/compression.h:677: undefined reference to LZ4_resetStreamHC'
/export/servers/temp/rocksdb/./util/compression.h:682: undefined reference to LZ4_loadDictHC' /export/servers/temp/rocksdb/./util/compression.h:687: undefined reference to LZ4_compress_HC_continue'
/export/servers/temp/rocksdb/./util/compression.h:693: undefined reference to LZ4_freeStreamHC' /usr/local/lib/librocksdb.a(block_based_table_builder.o): In function LZ4_Compress':
/export/servers/temp/rocksdb/./util/compression.h:560: undefined reference to `LZ4_loadDict'
collect2: error: ld returned 1 exit status

change build.sh
CGO_CFLAGS="-I/usr/local/include" CGO_LDFLAGS="-L/usr/local/lib -lrocksdb -lstdc++ -lm -lz -lbz2 -lzstd -llz4" go build -ldflags "-X main.CommitID=${CommitID} -X main.BranchName=${BranchName} -X 'main.BuildTime=${BuildTime}'" -o cfs-server
it can work

1. We should not have any items related to UMP in our open source release
2. We should rename the "util" folder to "utils"

It would be great to have a simple hugo website for CFS

I can have @lucperkins spin something up for the project

In the meta.json , "raftReplicatePort": "9094", the key must be change to "raftReplicaPort",so it can work

Provide information on how to:

• obtain, provide feedback (as bug reports or enhancements), and contribute to the software
• explain the contribution process (e.g., are pull requests used?)
• the requirements for acceptable contributions (e.g., a reference to any required coding standard).

as is indicated in CII Best Practices.

ML-powered policies to move around warm or cold objects to different storage devices

Currently there is no support to file name that does not use UTF-8 encoding.

Erasure coding of the cold extents of append-only large files can make storage space more efficient. Leveraging temporal locality to predict files and extents to be coded.

This internal name prefix should be removed in the open source release

cmd.go
master.json

./cmd -c meta.json

If no master in "masterAddrs" can be reached, cmd becomes unreponsive (no ctrl-c) and can only be killed with kill -9

In addition metaNode_error.log and metaNode_debug.log rapidly fill up until no more disk space is left.

maybe cmd can just quit if no master is available or try to reach a master for a certain time before it quits

In create volume API, the CFS create data partition for this volume when init volume??
In create data partition API， Does it mean expanding the volume?

• I could not find any documentation on how many datanodes(or disks?) need to be available.
• using the configuration files from chubaofs/docker/conf/ with different ips
• creating 2 masters, 3 metaservers, one datanode
• then running run_test.sh
• it shows create datapartition ok, but the partition is not created.

curl -v "http://192.168.0.131:17010/dataPartition/create?count=400&name=ltptest"

{"code":0,"msg":"success","data":" createDataPartition succeeeds. clusterLastTotalDataPartitions[0],vol[ltptest] has 0 data partitions previously and 0 data partitions now \n err[action[createDataPartition],clusterID[ltptest] vol[ltptest] Err:[cluster.go 550] no node set available for creating a data partition ]"}

log:
[ERROR] cluster.go:495: action[createDataPartition],clusterID[ltptest] vol[ltptest] Err:[cluster.go 550] no node set available for creating a data partition

curl -v "http://192.168.0.131:17010/admin/getCluster" | python -m json.tool

{
    "code": 0,
    "data": {
        "Applied": 23,
        "BadPartitionIDs": [],
        "DataNodeStatInfo": {
            "IncreasedGB": 0,
            "TotalGB": 7373,
            "UsedGB": 1949,
            "UsedRatio": "0.264"
        },
        "DataNodes": [
            {
                "Addr": "192.168.0.102:17310",
                "ID": 3,
                "IsWritable": true,
                "Status": true
            }
        ],
        "DisableAutoAlloc": false,
        "LeaderAddr": "192.168.0.131:17010",
        "MaxDataPartitionID": 0,
        "MaxMetaNodeID": 6,
        "MaxMetaPartitionID": 3,
        "MetaNodeStatInfo": {
            "IncreasedGB": 0,
            "TotalGB": 23,
            "UsedGB": 0,
            "UsedRatio": "0.004"
        },
        "MetaNodeThreshold": 0.75,
        "MetaNodes": [
            {
                "Addr": "192.168.0.119:17210",
                "ID": 4,
                "IsWritable": true,
                "Status": true
            },
            {
                "Addr": "192.168.0.127:17210",
                "ID": 2,
                "IsWritable": true,
                "Status": true
            },
            {
                "Addr": "192.168.0.137:17210",
                "ID": 5,
                "IsWritable": true,
                "Status": true
            }
        ],
        "Name": "ltptest",
        "VolStatInfo": [
            {
                "Name": "ltptest",
                "TotalGB": 30,
                "UsedGB": 0,
                "UsedRatio": "0.000"
            }
        ]
    },
    "msg": "success"
}

CNCF requires a code of conduct, you can use ours as a template

There is no distributed file system declared to be suitable for edge computing. In the field of edge computing, although resources are limited, business systems still have strong demand for middleware and databases. There are several factors that are considered in the field of edge computing:

• A single edge computing machine is less than 10 units.
• There are thousands of edge unit clusters, and it is difficult to maintain stable operation of these systems.

Possible solution

Strongly embrace the Cloud-Native
Specifically use the operator mode to run the distributed file system on the edge of self-maintenance

It may be possible to build and publish edge computing middleware and edge computing distributed file systems that solve edge computing scenarios.

Hope that chubaofs can do this work and incorporate the edge computing distributed file system requirements into chebaofs‘ Roadmap.

datanode.json

"disks": [
        "/mnt/btrfs/cha1:2474836480",
        "/doesnotexist:21474836480"
    ]

{1024 1048576}
Rlimit Final {1024000 1024000}
Fatal: failed to start the baud storage daemon - bad config file

1. the error message could actually say what the problem is
2. there is no error message in the dataNode_error.log

Fill out best practices badge for project:

https://bestpractices.coreinfrastructure.org/en

I am trying to build it on ubuntu 16.04
git clone https://github.com/facebook/rocksdb.git
apt install libgflags-dev libsnappy-dev zlib1g-dev libbz2-dev libzstd1-dev liblz4-dev
cd rocksdb && make clean && ROCKSDB_NO_FBCODE=1 make static_lib -j12 && make install

cd && export GOPATH=$HOME/go
go get github.com/chubaofs/chubaofs
cd ~/go/src/github.com/chubaofs/chubaofs/cmd && sh build.sh

github.com/chubaofs/chubaofs/vendor/github.com/tiglabs/raft/proto
../vendor/github.com/tiglabs/raft/proto/codec.go:275: undefined: sort.Slice
github.com/chubaofs/chubaofs/vendor/github.com/prometheus/common/expfmt
../vendor/github.com/prometheus/common/expfmt/text_create.go:456: undefined: bytes.ContainsAny
github.com/chubaofs/chubaofs/vendor/github.com/tecbot/gorocksdb

//usr/local/lib/librocksdb.so: undefined reference to dlopen'
//usr/local/lib/librocksdb.so: undefined reference to dlclose'
//usr/local/lib/librocksdb.so: undefined reference to dlerror'
//usr/local/lib/librocksdb.so: undefined reference to dlsym'
collect2: error: ld returned 1 exit status

Please hook in http://fossa.io for license scanning

CFS API need more information about server，The user cannot be sure that this API is provided by that service.
this is API for create volume：curl -v "http://127.0.0.1/admin/createVol?name=test&capacity=100&owner=cfs&mpCount=3"，but I can not sure which server support
this API

We need an authentication mechanism to address the following issues:

• Sometimes there are tens of thousands of clients accessing the same volume, and some STALE mounted clients are a potential threat to data safety.
• Some admin requests to master nodes are dangerous which requires authentication.

Therefore, we would like to introduce a new role called authnode to provide authentication capability.

It is important to clearify that assume we are in a safe network environment. So issues from an unsafe network environment such as hacking are out of the scope of this issue, and might be tracked by a new issue if needed.

It is a requirement for CNCF to have the project enforce DCO.

https://developercertificate.org/

It's good to highlight why CFS is better than other storage solutions.

You can look at the Prometheus page as an example on how to do this tastefully:
https://prometheus.io/docs/introduction/comparison/

I saw two types of Copyright headers:

Copyright 2018 The ChubaoFS Authors.

Copyright 2018 The Chubao Authors.

Add a file that describes how maintainers are added/removed

Here are some examples from CNCF projects:

https://github.com/coredns/coredns/blob/master/GOVERNANCE.md
https://github.com/envoyproxy/envoy/blob/master/GOVERNANCE.md
https://prometheus.io/governance

CFS will use Hole Punching for delete small file, but the macos do not support this API in systemcall

Some applications, e.g. image and short video services, do not need user-given file names or namespace hierarchy. It is enough to put an image and return a system-generated key.

Let's provide the independent 'key-file' API to data subsystem for atomic small file operations.

1、version 5.9.2
2、install rocksdb dependency；
3、make static_lib -j 'nproc' --can build faster
4、make install

Is there a public CI/build?

list MAINTAINERS/OWNERS who have commit access

panic: init UMP Monitor failed open /cha/warn/ump_metanode_tp.log: permission denied [recovered]
panic: init UMP Monitor failed open /cha/warn/ump_metanode_tp.log: permission denied

goroutine 1 [running]:
main.main.func1()
/root/go/src/github.com/chubaofs/chubaofs/cmd/cmd.go:112 +0xc7
panic(0x9c49c0, 0xc0006e4a40)
/usr/local/go/src/runtime/panic.go:522 +0x1b5
github.com/chubaofs/chubaofs/util/ump.InitUmp(0xc00053c060, 0x8, 0xc00053c0d0, 0xa)
/root/go/src/github.com/chubaofs/chubaofs/util/ump/ump.go:62 +0xee
main.main()
/root/go/src/github.com/chubaofs/chubaofs/cmd/cmd.go:139 +0x5ea

also :

panic: Failed to Open rocksDB! err:action[openRocksDB],err:IO error: No such file or directoryWhile mkdir if missing: /cha/Data/master/rocksdbstore: No such file or directory [recovered]
	panic: Failed to Open rocksDB! err:action[openRocksDB],err:IO error: No such file or directoryWhile mkdir if missing: /cha/Data/master/rocksdbstore: No such file or directory

the /cha/warn /cha/log /cha/Data directories do not exist when starting cmd.

ls -l
drw-r--r-- 2 owner owner 4096 Jun 20 11:57 warn/

some more friendly permissions
chmod a+x {warn,log}
mkdir -p /cha/Data/master
and
./cmd -c master.json seems to work

I think the directory structure must be created before starting cmd or the directories must be checked and created correctly when starting cmd

master.json

{
  "role": "master",
  ...
"logDir": "/cha/log/master",
"warnLogDir":"/cha/warn/",
...
}

The logo image showed on cellphone is discorded, please update

https://landscape.cncf.io/category=cloud-native-storage&format=card-mode

Chubao needs a lightweight distributed transaction mechanism to guarantee the consistency of cross-partition metadata update operations.

At the moment, there are no releases (i.e., git-tags)

How will CFS release new versions and tag them? Also how will they distribute them? (e.g, dockerhub, artifactory etc)

Running 3 masters on 3 different computers.
on each computer the cpu utilisation for cfs-server is around 50-70%
Is this normal?

top
  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
28946 root       20   0 1746172  62144  16792 S  71.5  0.8  35:24.51 cfs-server

git clone http://github.com/chubaofs/chubaofs.git
cd chubaofs/
make

build snappy...
build/build.sh: line 19: cmake: command not found
build snappy failed
Makefile:9: recipe for target 'build' failed
make: *** [build] Error 1