MinIO is a high performance object storage server compatible with Amazon S3 APIs
In a previous article, I mentioned minio as an S3 gateway between my system and backblaze b2. I was impressed by minio. So in this blog post, I would like to investigate the primary use of minio as an S3 storage provider!
Install Minio
Minio, is also software written in Golang. That means we can simple use the static binary executable in our machine.
Download
The latest release of minio is here:
curl -sLO https://dl.min.io/server/minio/release/linux-amd64/minio
chmod +x minio
Version
./minio version
$ ./minio version
Version: 2019-08-01T22:18:54Z
Release-Tag: RELEASE.2019-08-01T22-18-54Z
Commit-ID: c5ac901e8dac48d45079095a6bab04674872b28b
Operating System
Although we can use the static binary from minio’s site, I would propose to install minio through your distribution’s package manager, in Arch Linux is:
$ sudo pacman -S minio
this method, will also provide you, with a simple systemd service unit and a configuration file.
/etc/minio/minio.conf
# Local export path.
MINIO_VOLUMES="/srv/minio/data/"
# Access Key of the server.
# MINIO_ACCESS_KEY=Server-Access-Key
# Secret key of the server.
# MINIO_SECRET_KEY=Server-Secret-Key
# Use if you want to run Minio on a custom port.
# MINIO_OPTS="--address :9199"
Docker
Or if you like docker, you can use docker!
docker pull minio/minio
docker run -p 9000:9000 minio/minio server /data
Standalone
We can run minion as standalone
$ minio server /data
Create a test directory to use as storage:
$ mkdir -pv minio_data/
mkdir: created directory 'minio_data/'
$ /usr/bin/minio server ./minio_data/
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ You are running an older version of MinIO released 1 week ago ┃
┃ Update: Run `minio update` ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
Endpoint: http://192.168.1.3:9000 http://192.168.42.1:9000 http://172.17.0.1:9000 http://172.18.0.1:9000 http://172.19.0.1:9000 http://192.168.122.1:9000 http://127.0.0.1:9000
AccessKey: KYAS2LSSPXRZFH9P6RHS
SecretKey: qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Browser Access:
http://192.168.1.3:9000 http://192.168.42.1:9000 http://172.17.0.1:9000 http://172.18.0.1:9000 http://172.19.0.1:9000 http://192.168.122.1:9000 http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://192.168.1.3:9000 KYAS2LSSPXRZFH9P6RHS qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Object API (Amazon S3 compatible):
Go: https://docs.min.io/docs/golang-client-quickstart-guide
Java: https://docs.min.io/docs/java-client-quickstart-guide
Python: https://docs.min.io/docs/python-client-quickstart-guide
JavaScript: https://docs.min.io/docs/javascript-client-quickstart-guide
.NET: https://docs.min.io/docs/dotnet-client-quickstart-guide
Update Minio
okay, our package is from one week ago, but that’s okay. We can overwrite our package build (although not
recommended) with this:
$ sudo curl -sLo /usr/bin/minio https://dl.min.io/server/minio/release/linux-amd64/minio
again, NOT recommended.
Check version
minio version
Version: 2019-08-01T22:18:54Z
Release-Tag: RELEASE.2019-08-01T22-18-54Z
Commit-ID: c5ac901e8dac48d45079095a6bab04674872b28b
minio update
An alternative way, is to use the built-in update method:
$ sudo minio update
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ You are running an older version of MinIO released 5 days ago ┃
┃ Update: https://dl.min.io/server/minio/release/linux-amd64/minio ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
Update to RELEASE.2019-08-07T01-59-21Z ? [y/n]: y
MinIO updated to version RELEASE.2019-08-07T01-59-21Z successfully.
minio version
Version: 2019-08-07T01:59:21Z
Release-Tag: RELEASE.2019-08-07T01-59-21Z
Commit-ID: 930943f058f01f37cfbc2265d5f80ea7026ec55d
Run minio
run minion as standalone and localhost (not exposing our system to outside):
minio server --address 127.0.0.1:9000 ~/./minio_data/
output
$ minio server --address 127.0.0.1:9000 ~/./minio_data/
Endpoint: http://127.0.0.1:9000
AccessKey: KYAS2LSSPXRZFH9P6RHS
SecretKey: qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Browser Access:
http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://127.0.0.1:9000 KYAS2LSSPXRZFH9P6RHS qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Object API (Amazon S3 compatible):
Go: https://docs.min.io/docs/golang-client-quickstart-guide
Java: https://docs.min.io/docs/java-client-quickstart-guide
Python: https://docs.min.io/docs/python-client-quickstart-guide
JavaScript: https://docs.min.io/docs/javascript-client-quickstart-guide
.NET: https://docs.min.io/docs/dotnet-client-quickstart-guide
Web Dashboard
minio comes with it’s own web dashboard!
New Bucket
Let’s create a new bucket for testing purposes:
Minio Client
minio comes with it’s own minio client or mc
Install minio client
Binary Download
curl -sLO https://dl.min.io/client/mc/release/linux-amd64/mc
or better through your package manager:
sudo pacman -S minio-client
Access key / Secret Key
Now export our AK/SK in our enviroment
export -p MINIO_ACCESS_KEY=KYAS2LSSPXRZFH9P6RHS
export -p MINIO_SECRET_KEY=qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
minio host
or you can configure the minio server as a host:
./mc config host add myminio http://127.0.0.1:9000 KYAS2LSSPXRZFH9P6RHS qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
I prefer this way, cause I dont have to export keys every time.
List buckets
$ mc ls myminio
[2019-08-05 20:44:42 EEST] 0B bucket0001/
$ mc ls myminio/bucket0001
(empty)
List Policy
mc admin policy list myminio
$ mc admin policy list myminio
readonly
readwrite
writeonly
Credentials
If we do not want to get random Credentials every time, we can define them in our environment:
export MINIO_ACCESS_KEY=admin
export MINIO_SECRET_KEY=password
minio server --address 127.0.0.1:9000 .minio_data{1...10}
with minio client:
$ mc config host add myminio http://127.0.0.1:9000 admin password
mc: Configuration written to `/home/ebal/.mc/config.json`. Please update your access credentials.
mc: Successfully created `/home/ebal/.mc/share`.
mc: Initialized share uploads `/home/ebal/.mc/share/uploads.json` file.
mc: Initialized share downloads `/home/ebal/.mc/share/downloads.json` file.
Added `myminio` successfully.
mc admin config get myminio/ | jq .credential
$ mc admin config get myminio/ | jq .credential
{
"accessKey": "8RMC49VEC1IHYS8FY29Q",
"expiration": "1970-01-01T00:00:00Z",
"secretKey": "AY+IjQZomX6ZClIBJrjgxRJ6ugu+Mpcx6rD+kr13",
"status": "enabled"
}
s3cmd
Let’s configure s3cmd to use our minio data server:
$ sudo pacman -S s3cmd
Configure s3cmd
s3cmd --configure
$ s3cmd --configure
Enter new values or accept defaults in brackets with Enter.
Refer to user manual for detailed description of all options.
Access key and Secret key are your identifiers for Amazon S3. Leave them empty for using the env variables.
Access Key: KYAS2LSSPXRZFH9P6RHS
Secret Key: qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Default Region [US]:
Use "s3.amazonaws.com" for S3 Endpoint and not modify it to the target Amazon S3.
S3 Endpoint [s3.amazonaws.com]: http://127.0.0.1:9000
Use "%(bucket)s.s3.amazonaws.com" to the target Amazon S3. "%(bucket)s" and "%(location)s" vars can be used
if the target S3 system supports dns based buckets.
DNS-style bucket+hostname:port template for accessing a bucket [%(bucket)s.s3.amazonaws.com]:
Encryption password is used to protect your files from reading
by unauthorized persons while in transfer to S3
Encryption password:
Path to GPG program [/usr/bin/gpg]:
When using secure HTTPS protocol all communication with Amazon S3
servers is protected from 3rd party eavesdropping. This method is
slower than plain HTTP, and can only be proxied with Python 2.7 or newer
Use HTTPS protocol [Yes]: n
On some networks all internet access must go through a HTTP proxy.
Try setting it here if you can't connect to S3 directly
HTTP Proxy server name:
New settings:
Access Key: KYAS2LSSPXRZFH9P6RHS
Secret Key: qPZnIBJDe6GTRrUWcfdtKk7GPL4fGyqANDzJxkur
Default Region: US
S3 Endpoint: http://127.0.0.1:9000
DNS-style bucket+hostname:port template for accessing a bucket: %(bucket)s.s3.amazonaws.com
Encryption password:
Path to GPG program: /usr/bin/gpg
Use HTTPS protocol: False
HTTP Proxy server name:
HTTP Proxy server port: 0
Test access with supplied credentials? [Y/n] y
Please wait, attempting to list all buckets...
ERROR: Test failed: [Errno -2] Name or service not known
Retry configuration? [Y/n] n
Save settings? [y/N] y
Configuration saved to '/home/ebal/.s3cfg'
Test it
$ s3cmd ls
2019-08-05 17:44 s3://bucket0001
Distributed
Let’s make a more complex example and test the distributed capabilities of minio
Create folders
mkdir -pv .minio_data{1..10}
$ mkdir -pv .minio_data{1..10}
mkdir: created directory '.minio_data1'
mkdir: created directory '.minio_data2'
mkdir: created directory '.minio_data3'
mkdir: created directory '.minio_data4'
mkdir: created directory '.minio_data5'
mkdir: created directory '.minio_data6'
mkdir: created directory '.minio_data7'
mkdir: created directory '.minio_data8'
mkdir: created directory '.minio_data9'
mkdir: created directory '.minio_data10'
Start Server
Be-aware you have to user 3 dots (…) to enable erasure-code distribution (see below).
and start minio server like this:
minio server --address 127.0.0.1:9000 .minio_data{1...10}
$ minio server --address 127.0.0.1:9000 .minio_data{1...10}
Waiting for all other servers to be online to format the disks.
Status: 10 Online, 0 Offline.
Endpoint: http://127.0.0.1:9000
AccessKey: CDSBN216JQR5B3F3VG71
SecretKey: CE+ti7XuLBrV3uasxSjRyhAKX8oxtZYnnEwRU9ik
Browser Access:
http://127.0.0.1:9000
Command-line Access: https://docs.min.io/docs/minio-client-quickstart-guide
$ mc config host add myminio http://127.0.0.1:9000 CDSBN216JQR5B3F3VG71 CE+ti7XuLBrV3uasxSjRyhAKX8oxtZYnnEwRU9ik
Object API (Amazon S3 compatible):
Go: https://docs.min.io/docs/golang-client-quickstart-guide
Java: https://docs.min.io/docs/java-client-quickstart-guide
Python: https://docs.min.io/docs/python-client-quickstart-guide
JavaScript: https://docs.min.io/docs/javascript-client-quickstart-guide
.NET: https://docs.min.io/docs/dotnet-client-quickstart-guide
configure mc
$ ./mc config host add myminio http://127.0.0.1:9000 WWFUTUKB110NS1V70R27 73ecITehtG2rOF6F08rfRmbF+iqXjNr6qmgAvdb2
Added `myminio` successfully.
admin info
mc admin info myminio
$ mc admin info myminio
● 127.0.0.1:9000
Uptime: 3 minutes
Version: 2019-08-07T01:59:21Z
Storage: Used 25 KiB
Drives: 10/10 OK
Create files
Creating random files
for i in $(seq 10000) ;do echo $RANDOM > file$i ; done
and by the way, we can use mc to list our local files also!
$ mc ls file* | head
[2019-08-05 21:27:01 EEST] 6B file1
[2019-08-05 21:27:01 EEST] 5B file10
[2019-08-05 21:27:01 EEST] 5B file100
[2019-08-05 21:27:01 EEST] 6B file11
[2019-08-05 21:27:01 EEST] 6B file12
[2019-08-05 21:27:01 EEST] 6B file13
[2019-08-05 21:27:01 EEST] 6B file14
[2019-08-05 21:27:01 EEST] 5B file15
[2019-08-05 21:27:01 EEST] 5B file16
Create bucket
mc ls myminio
$ mc mb myminio/bucket0002
Bucket created successfully `myminio/bucket0002`.
$ mc ls myminio
[2019-08-05 21:41:35 EEST] 0B bucket0002/
Copy files
mc cp file* myminio/bucket0002/
be patient, even in a local filesystem, it will take a long time.
Erasure Code
copying from MinIO docs
you may lose up to half (N/2) of the total drives
MinIO shards the objects across N/2 data and N/2 parity drives
Here is the
$ du -sh .minio_data*
79M .minio_data1
79M .minio_data10
79M .minio_data2
79M .minio_data3
79M .minio_data4
79M .minio_data5
79M .minio_data6
79M .minio_data7
79M .minio_data8
79M .minio_data9
but what size did our files had?
$ du -sh files/
40M files
Very insteresting.
$ tree .minio_data*
Here is shorter list, to get an idea how objects are structured: minio_data_tree.txt
$ mc ls myminio/bucket0002 | wc -l
10000
Delete a folder
Let’s see how handles corrupted disks, but before that let’s keep a hash of our files:
md5sum file* > /tmp/files.before
now remove:
$ rm -rf .minio_data10
$ ls -la
total 0
drwxr-x--- 1 ebal ebal 226 Aug 15 20:25 .
drwx--x---+ 1 ebal ebal 3532 Aug 15 19:13 ..
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data1
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data2
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data3
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data4
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data5
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data6
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data7
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data8
drwxr-x--- 1 ebal ebal 40 Aug 15 20:25 .minio_data9
Notice that folder: minio_data10 is not there.
mc admin info myminio/
$ mc admin info myminio/
● 127.0.0.1:9000
Uptime: 6 days
Version: 2019-08-14T20:37:41Z
Storage: Used 57 MiB
Drives: 9/10 OK
This is the msg in minio server console:
API: SYSTEM()
Time: 20:23:50 EEST 08/15/2019
DeploymentID: 7852c1e1-146a-4ce9-8a05-50ad7b925fef
Error: unformatted disk found
endpoint=.minio_data10
3: cmd/prepare-storage.go:40:cmd.glob..func15.1()
2: cmd/xl-sets.go:212:cmd.(*xlSets).connectDisks()
1: cmd/xl-sets.go:243:cmd.(*xlSets).monitorAndConnectEndpoints()
Error: unformatted disk found
We will see that minio will try to create the disk/volume/folder in our system:
$ du -sh .minio_data*
79M .minio_data1
0 .minio_data10
79M .minio_data2
79M .minio_data3
79M .minio_data4
79M .minio_data5
79M .minio_data6
79M .minio_data7
79M .minio_data8
79M .minio_data9
Heal
Minio comes with a healing ability:
$ mc admin heal --recursive myminio/
$ du -sh .minio_data*
79M .minio_data1
79M .minio_data10
79M .minio_data2
79M .minio_data3
79M .minio_data4
79M .minio_data5
79M .minio_data6
79M .minio_data7
79M .minio_data8
79M .minio_data9
$ mc admin heal --recursive myminio/
◐ bucket0002/file9999
10,000/10,000 objects; 55 KiB in 58m21s
┌────────┬────────┬─────────────────────┐
│ Green │ 10,004 │ 100.0% ████████████ │
│ Yellow │ 0 │ 0.0% │
│ Red │ 0 │ 0.0% │
│ Grey │ 0 │ 0.0% │
└────────┴────────┴─────────────────────┘