Difference between revisions of "ViSUS Docker Deployment"

From
Jump to: navigation, search
(Before starting)
(Content of the image)
 
(42 intermediate revisions by the same user not shown)
Line 1: Line 1:
== Before starting ==
+
== Quick start ==
  
 
Start your Docker and login into your account.  
 
Start your Docker and login into your account.  
Line 5: Line 5:
  
 
To log in from your terminal:
 
To log in from your terminal:
 
+
<syntaxhighlight lang="bash">
<pre>docker login
+
docker login
 
Username: your_username
 
Username: your_username
Password: XXXX</pre>
+
Password: XXXX
 +
</syntaxhighlight>
  
 
Now pull the ViSUS image:
 
Now pull the ViSUS image:
  docker pull visus/mod_visus
+
<syntaxhighlight lang="bash">
 +
docker pull visus/visus
 +
</syntaxhighlight>
 +
 
 +
This image contains:
 +
* the ViSUS Server
 +
* configuration files (see [[ViSUS configuration file]])
 +
* a simple 2D web viewer
 +
* some tools to convert data (see [[ViSUS Convert]]).
 +
 
 +
Run the image:
 +
<syntaxhighlight lang="bash">
 +
  docker run -it --rm -p 8080:80 visus/visus
 +
</syntaxhighlight>
 +
 
 +
Check if the server is running:
 +
<syntaxhighlight lang="bash">
 +
curl -v "http://localhost:8080/mod_visus?action=list"
 +
</syntaxhighlight>
 +
 
 +
You will get a list of the current datasets on the server.
 +
 
 +
=== Content of the image ===
 +
 
 +
This image contains folders with configuration files and utilities.
 +
The folders of interest are under /visus:
 +
* '''config''', it contains the visus server.config (see [[ViSUS configuration file]])
 +
* '''apache2''', it contains the apache2 config files that you can use for security settings (see [[ViSUS Server]])
 +
* '''visus''', it contains visus tool (see [[ViSUS Convert]]) to convert data to IDX format
 +
 
 +
The user can map those folders locally and edit them adding a volume (with the Docker parameter ''-v''):
 +
 
 +
<syntaxhighlight lang="bash">
 +
docker run -it --rm -p 8080:80 -v /path/local_directory:/data/folder_in_container visus/visus
 +
</syntaxhighlight>
 +
 
 +
when you run you docker image.
 +
 
 +
=== How to customize the list of datasets on your server ===
 +
The list of dataset that the server exposes are contained in the [[ViSUS configuration file]].
 +
 
 +
If you have already a [[ViSUS configuration file]] that you want to use, you can run the container as following:
 +
<syntaxhighlight lang="bash">
 +
docker run -d -p 8080:80 -v $PWD/config:/visus/config visus/visus
 +
</syntaxhighlight>
 +
 
 +
and edit the '''visus.config''' file (see [[ViSUS configuration file]]).
 +
 
 +
'''Note''': The server needs to be able to read and write the configuration file, so make sure that your ''server.config'' has the right privileges. For example:
 +
<syntaxhighlight lang="bash">
 +
chmod 766 /path/to/your/server.config
 +
</syntaxhighlight>
 +
 
 +
Making the configuration file writable will allow to add datasets dynamically using the action ''add_dataset'' (see [[ViSUS Server]]).
 +
 
 +
If you also want to customize the apache security (e.g. change the users) you can run the command to mount also the apache2 directory as following:
 +
<syntaxhighlight lang="bash">
 +
docker run -d -p 8080:80 -v $PWD/config:/visus/config -v $PWD/apache2:/visus/apache2 visus/visus
 +
</syntaxhighlight>
 +
 
 +
=== Convert your data to IDX using the Docker deployment tools ===
 +
 
 +
The ViSUS tools executables (see [[ViSUS Convert]]) are contained in the container under <code>/usr/local/visus/bin</code>.
 +
 
 +
For example, if you want to use the tools to convert a raw dataset using your docker deployment you can do it as follow.
 +
 
 +
Mount a folder that contains your ''raw'' data and your converted ''idx'' data as follow:
 +
<syntaxhighlight lang="bash">
 +
docker run -d -p 8080:80 -v /full/path/to/your/raw/:/home/visus/datasets/raw -v /full/path/to/your/idx/:/home/visus/datasets/idx visus/visus
 +
</syntaxhighlight>
 +
 
 +
Check that your container is running correctly (using <code>docker ps</code>) and take note of the ''Container ID''.
 +
 
 +
Now run a bash session on this container using:
 +
<syntaxhighlight lang="bash">
 +
docker exec -it <id_from_docker_ps> /bin/bash
 +
</syntaxhighlight>
 +
 
 +
From this bash session you can use your ViSUS converter to convert your data (just mounted in the container) as described in [[ViSUS Convert]].
 +
 
 +
=== Preview your data with the Web Viewer ===
  
== Example of using precompiled mod_visus ==
+
You can preview and navigate you data on the server using the webviewer using your browser at the URL: http://localhost:8080
  
Create 'server.config' file (replace 2kbit1 with your dataset name):
+
Here is a picture of the viewer:
 +
[[File:2dwebviewer.png|300px]]
  
<pre>cat &lt;&lt;EOF &gt; server.config
+
With the commands on the bottom bar you can:
&lt;?xml version=&quot;1.0&quot; ?&gt;
+
* change server
&lt;visus&gt;
+
* set a slice (if using a 3D dataset)
  &lt;dataset name='2kbit1' url='file:///home/visus/dataset/2kbit1/visus.idx' permissions='public'/&gt;
+
* set a timestep
&lt;/visus&gt;
+
* set palettes and min-max range
EOF</pre>
+
* choose a field of the selected dataset
Create Dockerfile:
 
  
<pre>cat &lt;&lt;EOF &gt; Dockerfile
+
== ViSUS Server for NetCDF data ==
FROM visus/mod_visus
 
COPY server.config /home/visus/server.config
 
EOF</pre>
 
Build mod_visus docker image:
 
  
<pre>docker build -t visus/my-server .</pre>
+
An addition on-demand conversion service for NetCDF datasets is available in the "ondemand" Docker container.
Run the image mounting a host volume:
 
  
<pre>docker run -it --rm -p 8080:80 --volume=&quot;c:/visus_dataset/2kbit1:/home/visus/dataset/2kbit1&quot; visus/my-server </pre>
+
To try it, pull the ViSUS on-demand image:
Test if it works:
+
<syntaxhighlight lang="bash">
 +
docker pull visus/ondemand
 +
</syntaxhighlight>
  
<pre>wget &quot;http://localhost:8080/mod_visus?action=readdataset&amp;dataset=2kbit1&quot;</pre>
+
This container provides three services:
== Example of creating a test david_subsampled server ==
+
* the ViSUS Server
 +
* conversion service (i.e., currently limited to NetCDF datasets)
 +
* a simple Web viewer
  
Create 'server.config' file:
+
=== How to run the Containter ===
  
<pre>cat &lt;&lt;EOF &gt; server.config
+
You need to specify few simple settings:
&lt;?xml version=&quot;1.0&quot; ?&gt;
+
* the port you want to use for the conversion service (internal 80)
&lt;visus&gt;
+
* the port you want to use for the webviewer and data server (internal 42299)
  &lt;dataset name='david_subsampled' url='file:///home/visus/dataset/david_subsampled/visus.idx' permissions='public'/&gt;
+
* the folder containing the files to serve/convert (i.e., NetCDF files), this will be mapped to "/data/xml" in the container
&lt;/visus&gt;
+
* a folder to use for caching the multiresolution data for streaming, this will be mapped to "/data/idx" in the container
EOF</pre>
 
Create Dockerfile:
 
  
<pre>cat &lt;&lt;EOF &gt; Dockerfile
+
Quick start with some demo data:
FROM visus/mod_visus
+
<syntaxhighlight lang="bash">
COPY server.config /home/visus/server.config
+
# create directory for your input data
RUN set -x \
+
mkdir -p $PWD/nc
  &amp;&amp; apt update \
+
# create directory for your cache
  &amp;&amp; apt install -y curl \
+
mkdir -p $PWD/idx
  &amp;&amp; cd /home/visus/dataset \
 
  &amp;&amp; curl http://atlantis.sci.utah.edu/download/david_subsampled.tar.gz -o temp.tar.gz \
 
  &amp;&amp; tar xvzf temp.tar.gz \
 
  &amp;&amp; rm -f temp.tar.gz \
 
  &amp;&amp; chown -R www-data /home/visus/dataset \
 
  &amp;&amp; chmod -R a+rX  /home/visus/dataset 
 
EOF</pre>
 
Build mod_visus docker image:
 
  
<pre>docker build -t visus/david_subsampled .</pre>
+
# download a sample NetCDF dataset
Run the image mounting a host volume:
+
wget https://www.dropbox.com/s/jnhptd5vswaurem/test2d.nc -O ./nc/test2d.nc
 +
</syntaxhighlight>
  
<pre>docker run -it --rm -p 8080:80 visus/david_subsampled </pre>
+
Run the container:
Test if it works:
+
<syntaxhighlight lang="bash">
 +
docker run -it -p 80:80 -p 42299:42299 -v $PWD/nc:/data/xml -v $PWD/idx:/data/idx visus/ondemand
 +
</syntaxhighlight>
  
<pre>wget &quot;http://localhost:8080/mod_visus?action=readdataset&amp;dataset=david_subsampled&quot;</pre>
+
In this run command we mapped:
== Some tricks ==
+
* port 80, used in the container for the data streaming and webviewer services
 +
* port 42299, used in the container for the NetCDF data conversion service
  
To debug mod_visus (example: to check Apache log files):
+
=== Visualize/convert your data on the server ===
  
<pre>docker run -ti -p 8080:80 --entrypoint=/bin/bash visus/mod_visus -s
+
Go to the On-demand test page: [http://localhost/ondemand.php http://localhost/ondemand.php].  
/usr/local/bin/httpd-foreground.sh</pre>
 
To run an interactive 'clean' ubuntu (or whatever):
 
  
<pre>docker run -ti -p 8080:80 --entrypoint=/bin/bash ubuntu:16.04 -s</pre>
+
Note: replace "localhost" with your server name and port if you are running with custom settings.
If you want to remove all old docker images and containers (and you know what you are doing!)
 
  
<pre>docker rm  -f $(docker ps  -a -q)
+
In this page you will find the list of *.nc contained in the folder you provided. You can click on the corresponding "View" button to visualize your data. Note: if the dataset has not been already converted the request could take a few second since it will be converting on-demand.
docker rmi -f $(docker images -q)</pre>
 

Latest revision as of 17:55, 23 January 2019

Quick start

Start your Docker and login into your account. If you don't have installed follow these instructions: Docker Documentation

To log in from your terminal:

docker login
Username: your_username
Password: XXXX

Now pull the ViSUS image:

docker pull visus/visus

This image contains:

Run the image:

docker run -it --rm -p 8080:80 visus/visus

Check if the server is running:

curl -v "http://localhost:8080/mod_visus?action=list"

You will get a list of the current datasets on the server.

Content of the image

This image contains folders with configuration files and utilities. The folders of interest are under /visus:

  • config, it contains the visus server.config (see ViSUS configuration file)
  • apache2, it contains the apache2 config files that you can use for security settings (see ViSUS Server)
  • visus, it contains visus tool (see ViSUS Convert) to convert data to IDX format

The user can map those folders locally and edit them adding a volume (with the Docker parameter -v):

docker run -it --rm -p 8080:80 -v /path/local_directory:/data/folder_in_container visus/visus

when you run you docker image.

How to customize the list of datasets on your server

The list of dataset that the server exposes are contained in the ViSUS configuration file.

If you have already a ViSUS configuration file that you want to use, you can run the container as following:

docker run -d -p 8080:80 -v $PWD/config:/visus/config visus/visus

and edit the visus.config file (see ViSUS configuration file).

Note: The server needs to be able to read and write the configuration file, so make sure that your server.config has the right privileges. For example:

chmod 766 /path/to/your/server.config

Making the configuration file writable will allow to add datasets dynamically using the action add_dataset (see ViSUS Server).

If you also want to customize the apache security (e.g. change the users) you can run the command to mount also the apache2 directory as following:

docker run -d -p 8080:80 -v $PWD/config:/visus/config -v $PWD/apache2:/visus/apache2 visus/visus

Convert your data to IDX using the Docker deployment tools

The ViSUS tools executables (see ViSUS Convert) are contained in the container under /usr/local/visus/bin.

For example, if you want to use the tools to convert a raw dataset using your docker deployment you can do it as follow.

Mount a folder that contains your raw data and your converted idx data as follow:

docker run -d -p 8080:80 -v /full/path/to/your/raw/:/home/visus/datasets/raw -v /full/path/to/your/idx/:/home/visus/datasets/idx visus/visus

Check that your container is running correctly (using docker ps) and take note of the Container ID.

Now run a bash session on this container using:

docker exec -it <id_from_docker_ps> /bin/bash

From this bash session you can use your ViSUS converter to convert your data (just mounted in the container) as described in ViSUS Convert.

Preview your data with the Web Viewer

You can preview and navigate you data on the server using the webviewer using your browser at the URL: http://localhost:8080

Here is a picture of the viewer: 2dwebviewer.png

With the commands on the bottom bar you can:

  • change server
  • set a slice (if using a 3D dataset)
  • set a timestep
  • set palettes and min-max range
  • choose a field of the selected dataset

ViSUS Server for NetCDF data

An addition on-demand conversion service for NetCDF datasets is available in the "ondemand" Docker container.

To try it, pull the ViSUS on-demand image:

docker pull visus/ondemand

This container provides three services:

  • the ViSUS Server
  • conversion service (i.e., currently limited to NetCDF datasets)
  • a simple Web viewer

How to run the Containter

You need to specify few simple settings:

  • the port you want to use for the conversion service (internal 80)
  • the port you want to use for the webviewer and data server (internal 42299)
  • the folder containing the files to serve/convert (i.e., NetCDF files), this will be mapped to "/data/xml" in the container
  • a folder to use for caching the multiresolution data for streaming, this will be mapped to "/data/idx" in the container

Quick start with some demo data:

# create directory for your input data
mkdir -p $PWD/nc
# create directory for your cache
mkdir -p $PWD/idx

# download a sample NetCDF dataset
wget https://www.dropbox.com/s/jnhptd5vswaurem/test2d.nc -O ./nc/test2d.nc

Run the container:

docker run -it -p 80:80 -p 42299:42299 -v $PWD/nc:/data/xml -v $PWD/idx:/data/idx visus/ondemand

In this run command we mapped:

  • port 80, used in the container for the data streaming and webviewer services
  • port 42299, used in the container for the NetCDF data conversion service

Visualize/convert your data on the server

Go to the On-demand test page: http://localhost/ondemand.php.

Note: replace "localhost" with your server name and port if you are running with custom settings.

In this page you will find the list of *.nc contained in the folder you provided. You can click on the corresponding "View" button to visualize your data. Note: if the dataset has not been already converted the request could take a few second since it will be converting on-demand.