building the workload from a local image about cloudsuite

[neo-apz]

What you did instead of docker create was supposed to be done instead of step 5 not docker create. So in step 5 instead of

docker build data-analytics/dataset

Use this:

docker build -t data-analytics:dataset ./data-analytics/dataset

By using this, your built image has the required tag, data-analytics:dataset, which is then can be used in other commands like docker create command.

BTW, in step 4, the ubuntu version must be 14.04 not 10.04.

[mahmoodn]

I successfully ran the create command. Now, I have to start the master. Two commands should be used

$ docker pull cloudsuite/data-analytics:master
$ docker run -d -t --dns 127.0.0.1 -P --name master -h master.cloudsuite.com --volumes-from data cloudsuite/data-analytics:master

Instead, I tried

mahmood@orca:cloudsuite_v3$ mkdir -p data-analytics/master
mahmood@orca:cloudsuite_v3$ vim data-analytics/master/Dockerfile

(the Dockerfile content is the same as this)

mahmood@orca:cloudsuite_v3$ mkdir -p data-analytics/master
mahmood@orca:cloudsuite_v3$ docker build data-analytics/master
Sending build context to Docker daemon 4.096 kB
Step 1 : FROM cloudsuite/data-analytics:base
Pulling repository docker.io/cloudsuite/data-analytics_

As you can see, it again tries to pull. What should I do?
Also, should I download files and folders of master from the github and move them to my local host?

[neo-apz]

The base image (check FROM statement) of this image is data-analytics:base. So you have to first build that one and then try to build data-analytics:master.

BTW, you can download the whole repository as an archive file. Then at root the archived folder, you can easily build each image using this command:

docker build -t $DH_REPO:$IMG_TAG $DF_PATH

For instance for data-analytics the three variables are:

DH_REPO="cloudsuite/data-analytics" IMG_TAG="base" DF_PATH="./benchmarks/data-analytics/base"
DH_REPO="cloudsuite/data-analytics" IMG_TAG="dataset" DF_PATH="./benchmarks/data-analytics/dataset"
DH_REPO="cloudsuite/data-analytics" IMG_TAG="master" DF_PATH="./benchmarks/data-analytics/master"
DH_REPO="cloudsuite/data-analytics" IMG_TAG="slave" DF_PATH="./benchmarks/data-analytics/slave"

For more information about the docker commands and the building process of Docker images, you can take a look at the docker documentation.

[mahmoodn]

1- FYI, there is no information about building the base/ in the data-analytics guide.
2- Should I use build -t (similar to the dataset) for all folders? Or that -t is used for dataset only?
3- The serf package from bintray.com is no longer available. Is that the same as serf.googlecode.com? Please provide us the package to prevent any version mismatch.

[aledaglis]

First off, I'd like to mention that the vast majority of this thread is about Docker, not CloudSuite. We offer CloudSuite 3.0 with Docker just to ease the process of deploying it. If, for any reason, you want to follow your own setup and not follow the instructions, you'll have to become more familiar with Docker, and this is not something our documentation/support aspires to cover. If this is an overhead you'd prefer to avoid, you can follow the Dockerfiles as instructions for a bare metal setup (w/o Docker).

To answer your Q's:
1- Following what I mentioned above, there is indeed no special information about building the base, as the way you are setting it up is not our intended use case. This means you have to follow the commands in the Dockerfiles on your own. In this particular case, the Dockerfile specifies that the base image is required (as implied by the FROM keyword).
2- I'm not sure I understand the question. The -t option of the build command is used to tag the image you're building, in order to make it easier to refer to it.
3- Thanks for pointing that out. I opened a separate issue for that to be updated. You can find the package that is used by the Dockerfile here.

[neo-apz]

@mahmoodn Take a look at this. It may help you solving your problem.