S3 gateway for cluster is not responding

cn (ceph-nano)

The project

cn is a little program written in Go that helps you interact with the S3 API by providing a REST S3 compatible gateway. The target audience is developers building their applications on Amazon S3. It is also an exciting tool to showcase Ceph Rados Gateway S3 compatibility. This is brought to you by the power of Ceph and Containers. Under the hood, cn runs a Ceph container and exposes a Rados Gateway. For convenience, cn also comes with a set of commands to work with the S3 gateway. Before you ask "why not using s3cmd instead?", then you will be happy to read that internally cn uses s3cmd and act as a wrapper around the most commonly used commands. Also, keep in mind that the CLI is just for convenience, and the primary use case is you developing your application directly on the S3 API.

Build

You can build cn by using make. Be sure dep is installed:

$ go get github.com/golang/dep/cmd/dep

Then, add ~/go/bin to your $PATH:

$ export PATH=$PATH:~/go/bin

Build cn:

$ make
rm -f cn cn &>/dev/null || true
dep ensure
GOOS=linux GOARCH=amd64 go build -i -ldflags="-X main.version=cea247c-dirty -X main.tag=devel -X main.branch=guits-doc_build" -o cn-devel-cea247c-dirty-linux-amd64 main.go
ln -sf "cn-devel-cea247c-dirty-linux-amd64" cn

Once the build is done, you should have a symlink cn pointing to the binary that just got built:

$ ls -l
total 10692
-rw-rw-r--. 1 guits guits    15292 20 nov.  22:03 ceph-nano-logo-vertical.jpg
drwxrwxr-x. 2 guits guits     4096 20 nov.  22:03 cmd
lrwxrwxrwx. 1 guits guits       34 20 nov.  22:27 cn -> cn-devel-cea247c-dirty-linux-amd64
-rwxrwxr-x. 1 guits guits 10881196 20 nov.  22:27 cn-devel-cea247c-dirty-linux-amd64

Installation

cn relies on Docker so it must be installed on your machine. If you're not running a Linux workstation you can install Docker for Mac.

Once Docker is installed you're ready to start. Open your terminal and download the cn binary.

macOS:

curl -L https://github.com/ceph/cn/releases/download/v2.3.1/cn-v2.3.1-darwin-amd64 -o cn && chmod +x cn

Linux amd64:

curl -L https://github.com/ceph/cn/releases/download/v2.3.1/cn-v2.3.1-linux-amd64 -o cn && chmod +x cn

Linux arm64:

curl -L https://github.com/ceph/cn/releases/download/v2.3.1/cn-v2.3.1-linux-arm64 -o cn && chmod +x cn

Test it out

$ ./cn
Ceph Nano - One step S3 in container with Ceph.

                  *(((((((((((((
                (((((((((((((((((((
              ((((((((*     ,(((((((*
             ((((((             ((((((
            *((((,               ,((((/
            ((((,     ((((((/     *((((
            ((((     (((((((((     ((((
            /(((     (((((((((     ((((
             (((.     (((((((     /(((/
              (((                *((((
              .(((              (((((
         ,(((((((*             /(((
          .(((((  ((( (/  //   (((
                 /(((.  /(((((  /(((((
                        .((((/ (/

Usage:
  cn [command]

Available Commands:
  cluster      Interact with a particular Ceph cluster
  s3           Interact with a particular S3 object server
  image        Interact with cn's container image(s)
  version      Print the version of cn
  kube         Outputs cn kubernetes template (cn kube > kube-cn.yml)
  update-check Print cn current and latest version number
  flavors      Interact with flavors
  completion   Generates bash completion scripts

Flags:
  -h, --help   help for cn

Use "cn [command] --help" for more information about a command.

Get started

Start the program with a working directory /tmp, the initial start might take a few minutes since we need to download the container image:

$ ./cn cluster start -d /tmp my-first-cluster
Running ceph-nano...
The container image is not present, pulling it.
This operation can take a few minutes......................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

Endpoint: http://10.36.116.164:8000
Dashboard: http://10.36.116.164:5001
Access key is: 9ZU1QBYX13KPLXXDDCY2
Secret key is: nthNG1xb7ta5IDKiJKM8626pQitqsalEo0ta7B9E
Working directory: /usr/share/ceph-nano

Selecting the cluster flavor

Is it possible to select the cluster flavor by using the -f option on the command line.

$ ./cn cluster start mycluster -f huge

The full documentation of flavors can be found here

Your first S3 bucket

Create a bucket with cn:

$ ./cn s3 mb my-first-cluster my-buc
Bucket 's3://my-buc/' created

$ ./cn s3 put my-first-cluster /etc/passwd my-buc
upload: '/tmp/passwd' -> 's3://my-buc/passwd'  [1 of 1]
 5925 of 5925   100% in    1s     4.57 kB/s  done

Multi-cluster support

cn can manage any number of clusters on your local machine:

$ ./cn cluster ls
+------+---------+-------------------------------------------------------------------------------------+----------------+--------------------------------+---------+
| NAME | STATUS  | IMAGE                                                                               | IMAGE RELEASE  | IMAGE CREATION TIME            | FLAVOR  |
+------+---------+-------------------------------------------------------------------------------------+----------------+--------------------------------+---------+
| d    | running | ceph/daemon:latest                                                                  | master-77e3d8d | 2018-04-05T15:01:40.323603472Z | default |
| b    | running | ceph/daemon@sha256:369867e450ccdea9bcea7f54e97ed8b2cb1a0437fbef658d2d01fce2b8a2c648 | master-5f44af9 | 2018-03-30T21:08:31.117367166Z | medium  |
+------+---------+-------------------------------------------------------------------------------------+----------------+--------------------------------+---------+

List Ceph container images available

cn can list the available Ceph container images, the default output shows the 100 first images:

$ ./cn image ls
ceph/daemon:latest-mimic
ceph/daemon:latest-luminous
ceph/daemon:latest-master
ceph/daemon:master-0b3eb04-mimic-centos-7
ceph/daemon:master-0b3eb04-luminous-centos-7
ceph/daemon:master-0b3eb04-luminous-opensuse-42.3-x86_64
ceph/daemon:master-0b3eb04-master-centos-7-x86_64
ceph/daemon:master-0b3eb04-luminous-centos-7-x86_64
ceph/daemon:master-0b3eb04-mimic-centos-7-x86_64
[...]

Using images aliases

The image option (-i) support aliases to simply the command line. It is possible to list the aliases by running the image show-aliases command as per below :

$ ./cn image show-aliases
+----------+--------------------------------------------------+
| ALIAS    | IMAGE_NAME                                       |
+----------+--------------------------------------------------+
| mimic    | ceph/daemon:latest-mimic                         |
| luminous | ceph/daemon:latest-luminous                      |
| redhat   | registry.access.redhat.com/rhceph/rhceph-3-rhel7 |
+----------+--------------------------------------------------+

Aliases can be use in place of the traditional image name as per the following example:

$ ./cn cluster start mycluster -i mimic

It is also possible to create new aliases as detailed here

Enable mgr dashboard

TODO: This is a temporary hack to enable the manager dashboard

Currently cn does not expose a port for the mgr dashboard. It only exposes port 8000 for S3 API, and port 5000 for Sree - S3 web client. To expose also the mgr dashboard port we currently have to do some hacks.

This section will guide you how to manually commit a new image and then run a new container with the desired expose ports.

Commit a copy of the docker image:

./cn cluster start temp -d /tmp
docker commit ceph-nano-temp ceph-nano
./cn cluster purge temp --yes-i-am-sure

Run the container:

docker run -dt --name cn -p 8080:8080 -p 5000:5000 -p 8000:8000 ceph-nano

Enable dashboard:

(Note: 'enable dashboard' command will cause the container to exit, so need to start it after)

docker exec cn ceph config set mgr mgr/dashboard/ssl false
docker exec cn ceph config set mgr mgr/dashboard/server_addr 0.0.0.0
docker exec cn ceph config set mgr mgr/dashboard/server_port 8080
docker exec cn ceph mgr module enable dashboard
until docker exec cn ceph; do docker start cn; sleep 1; done # wait for the services to start
docker exec cn ceph dashboard set-login-credentials nano nano

Note that the Object Gateway tab in the dashboard is not enabled yet, so run the following to enable RGW dashboard:

RGW_USER=$(docker exec cn radosgw-admin user create --uid=rgw --display-name=rgw --system)
RGW_ACCESS=$(echo $RGW_USER | awk '{ for (i=1;i<=NF;++i) if ($i ~ /access_key/) { split($(i+1),a,"\""); print a[2] } }')
RGW_SECRET=$(echo $RGW_USER | awk '{ for (i=1;i<=NF;++i) if ($i ~ /secret_key/) { split($(i+1),a,"\""); print a[2] } }')
docker exec cn ceph dashboard set-rgw-api-access-key "$RGW_ACCESS"
docker exec cn ceph dashboard set-rgw-api-secret-key "$RGW_SECRET"
docker exec cn ceph dashboard set-rgw-api-host 127.0.0.1
docker exec cn ceph dashboard set-rgw-api-port 8000
docker exec cn ceph dashboard set-rgw-api-scheme http
docker exec cn ceph dashboard set-rgw-api-user-id rgw

The dashboard should now be accessible:

Open http://127.0.0.1:8080
Login with user nano and password nano

Troubleshooting - verify that your config dump should look like this:

$ docker exec cn ceph config dump
WHO   MASK LEVEL   OPTION                           VALUE                                                        RO 
  mgr      unknown mgr/dashboard/RGW_API_ACCESS_KEY ********************                                         *  
  mgr      unknown mgr/dashboard/RGW_API_HOST       127.0.0.1                                                    *  
  mgr      unknown mgr/dashboard/RGW_API_PORT       8000                                                         *  
  mgr      unknown mgr/dashboard/RGW_API_SCHEME     http                                                         *  
  mgr      unknown mgr/dashboard/RGW_API_SECRET_KEY ****************************************                     *  
  mgr      unknown mgr/dashboard/RGW_API_USER_ID    rgw                                                          *  
  mgr      unknown mgr/dashboard/password           ************************************************************ *  
  mgr      unknown mgr/dashboard/server_addr        0.0.0.0                                                      *  
  mgr      unknown mgr/dashboard/server_port        8080                                                         *  
  mgr      unknown mgr/dashboard/ssl                false                                                        *  
  mgr      unknown mgr/dashboard/username           nano                                                         *

	// getSeLinuxStatus gets SeLinux status
	func getSeLinuxStatus() string {
	testBinaryExist("getenforce")

	out, err := exec.Command("getenforce").Output()
	if err != nil {
	log.Fatal(err)
	}
	return string(out)
	}

	// applySeLinuxLabel checks if SeLinux is installed and set to Enforcing,
	// we relabel our workingDirectory to allow the container to access files in this directory
	func applySeLinuxLabel(dir string) {
	testBinaryExist("getenforce")

	selinuxStatus := getSeLinuxStatus()
	lines := strings.Split(selinuxStatus, "\n")
	for _, l := range lines {
	if len(l) <= 0 {
	// Ignore empty line.
	continue
	}
	if l == "Enforcing" {
	meUserName, meID := whoAmI()
	if meID != "0" {
	log.Fatal("Hey " + meUserName + "! Run me as 'root' so I can apply the right SeLinux label on " + dir)
	}
	if _, err := os.Stat(dir); os.IsNotExist(err) {
	os.Mkdir(dir, 0755)
	}
	testBinaryExist("chcon")
	cmd := "chcon " + " -Rt" + " svirt_sandbox_file_t " + dir
	_, err := exec.Command("chcon", "-Rt", "svirt_sandbox_file_t", dir).Output()
	if err != nil {
	log.Fatal(err)
	}
	log.Println("Executing: " + cmd)
	}
	}
	}

ceph / cn Goto Github PK

cn's Introduction

cn (ceph-nano)

The project

Table of contents

Build

Installation

Get started

Selecting the cluster flavor

Your first S3 bucket

Multi-cluster support

List Ceph container images available

Using images aliases

Enable mgr dashboard

TODO: This is a temporary hack to enable the manager dashboard

Commit a copy of the docker image:

Run the container:

Enable dashboard:

Note that the Object Gateway tab in the dashboard is not enabled yet, so run the following to enable RGW dashboard:

The dashboard should now be accessible:

Troubleshooting - verify that your config dump should look like this:

cn's People

Contributors

Stargazers

Watchers

Forkers

cn's Issues

Recommend Projects

Recommend Topics

Recommend Org