You should find h5serv quite easy to setup. The server (based on Python Tornado) isself-contained, so you will not need to setup Apache or other web server software to utilizeh5serv.
This guide from a UC Irvine Computer Science professor's page is a very well-written, detailed, and Mac-specfic description of all the steps involved in installing GDB, creating a certificate, signing GDB using that certificate, and finally configuring Eclipse. I have mac OS High Sierra 10.13.4, Eclipse 4.7.3a Now I can debug in Eclipse as usual in C language. I have tried disabling csrutil completely and used both macports and brew to install GDB 8.0.1. I'm trying to debug on CLion; GDB starts but freezes right after it starts and eventually I get the message.
Prerequisites¶
A computer running a 64-bit version of Windows, Mac OS X, or Linux.
You will also need the following Python packages:
- Python 2.7 or later
- NumPy 1.10.4 or later
- h5py 2.5 or later
- tornado 4.0.2 or later
- watchdog 0.8.3 or later
- requests 2.3 or later (for client tests)
If you are not familiar with installing Python packages, the easiest route is touse a package manager such as Anaconda (as described below).
If you have a git client installed on your system, you can directly download the h5servsource from GitHub: gitclone--recursivehttps://github.com/HDFGroup/h5serv.git.Otherwise, you can download a zip file of the source from GitHub (as described below).
Installing on Windows¶
Anaconda from Continuum Analytics can be used to easily manage the package dependenciesneeded for HDF Server.
In a browser go to: http://continuum.io/downloads and click the 'Windows 64-bitPython 2.7 Graphical Installer' button.
Install Anaconda using the default options.
Once Anaconda is installed select 'Anaconda Command Prompt' from the start menu.
In the command window that appears, create a new anaconda environment using the following command:condacreate-nh5servpython=2.7h5pytornadorequestspytz
Answer ‘y' to the prompt, and the packages will be fetched.
In the same command window, run: activateh5serv
Install the watchdog package (this is currently not available through Anaconda):pipinstallwatchdog
Download the hdf5-json project: gitclonehttps://github.com/HDFGroup/hdf5-json.git .Alternatively, in a browser go to: https://github.com/HDFGroup/hdf5-json and click the'Download ZIP' button (right side of page). Download the zip file and extract tothe destination directory of your choice.
Next, cd to the hdf5-json folder and run: pythonsetup.pyinstall.
Download the h5serv project: gitclonehttps://github.com/HDFGroup/h5serv.git .Alternatively, download the source zip as described above.
Next, in the command window, cd to the folder you extracted the source files to.
From here cd to 'h5serv-master/server' (or just server if you used git).
Run: python app.pyYou should see the output: 'Starting event loop on port: 5000'.
You may then see a security alert: 'Windows Firewall has blocked some features of thisprogram'. Click 'Allow access' with the default option (Private network access).
At this point the server is running, waiting on any requests being sent to port 5000.Go to the 'verification' section below to try out the service.
Installing on Linux/Mac OS X¶
Anaconda from Continuum Analytics can be used to easily manage the package dependenciesneeded for HDF Server.
In a browser go to: http://continuum.io/downloads and click the 'Mac OS X 64-bitPython 2.7 Graphical Installer' button for Mac OS X or: 'Linux 64-bit Python 2.7'.
Install Anaconda using the default options.
Once Anaconda is installed, open a new shell and run the following on the command line:
condacreate-nh5servpython=2.7h5pytornadorequestspytz
Answer ‘y' to the prompt, and the packages will be fetched.
Install the watchdog package (this is currently not available through Anaconda):pipinstallwatchdog
Deck the balls mac os. In the same shell, run: sourceactivateh5serv
Download the hdf5-json project: gitclonehttps://github.com/HDFGroup/hdf5-json.git .Alternatively, in a browser go to: https://github.com/HDFGroup/hdf5-json and click the'Download ZIP' button (right side of page). Download the zip file and extract tothe destination directory of your choice.
Next, cd to the hdf5-json folder and run: pythonsetup.pyinstall.
Download the h5serv project: gitclonehttps://github.com/HDFGroup/h5serv.git .Alternatively, download the source zip as described above.
Next, cd to the h5serv folder.
From here cd to 'server' (or 'h5serv-master/server' if you extracted from ZIP file).
Run: pythonapp.pyYou should see the output: 'Starting event loop on port: 5000'.
At this point the server is running, waiting on any requests being sent to port 5000.Go to the 'verification' section below to try out the service.
Verification¶
To verify that h5serv was installed correctly, you can run the test suite includedwith the installation.
Open a new shell (on Windows, run 'Annaconda Command Prompt' from the start menu).
In this shell, run the following commands:
- source activate h5serv (just: activate h5serv on Windows)
- cd
- cd test
- python testall.py
All tests should report OK.
Server Configuration¶
The file h5serv/server/config.py provides several configuration options that can beused to customize h5serv. Each of the options can be changed by:
- Changing the value in the config.py file and re-starting the service.
- Passing a command line option to app.py on startup. E.g.
pythonapp.py--port=7253 - Setting an environment variable with the option name in upper case. E.g.
exportPORT=5000;pythonapp.py
The config options are:
port¶
The port that h5serv will listen on. Change this if 5000 conflicts with another service.
Default: 5000
debug¶
If True the server will report debug info (e.g. a stack trace) to the requester onerror. If False, just the status code and message will be reported.
Default: True
datapath¶
A path indicating the directory where HDF5 files will be be stored.
Note: Any HDF5 file content that you put in this directory will be exposed via theserver REST api (unless the domain's ACL is configured to prevent public access, see:./AclOps).
If you have a git client installed on your system, you can directly download the h5servsource from GitHub: gitclone--recursivehttps://github.com/HDFGroup/h5serv.git.Otherwise, you can download a zip file of the source from GitHub (as described below).
Installing on Windows¶
Anaconda from Continuum Analytics can be used to easily manage the package dependenciesneeded for HDF Server.
In a browser go to: http://continuum.io/downloads and click the 'Windows 64-bitPython 2.7 Graphical Installer' button.
Install Anaconda using the default options.
Once Anaconda is installed select 'Anaconda Command Prompt' from the start menu.
In the command window that appears, create a new anaconda environment using the following command:condacreate-nh5servpython=2.7h5pytornadorequestspytz
Answer ‘y' to the prompt, and the packages will be fetched.
In the same command window, run: activateh5serv
Install the watchdog package (this is currently not available through Anaconda):pipinstallwatchdog
Download the hdf5-json project: gitclonehttps://github.com/HDFGroup/hdf5-json.git .Alternatively, in a browser go to: https://github.com/HDFGroup/hdf5-json and click the'Download ZIP' button (right side of page). Download the zip file and extract tothe destination directory of your choice.
Next, cd to the hdf5-json folder and run: pythonsetup.pyinstall.
Download the h5serv project: gitclonehttps://github.com/HDFGroup/h5serv.git .Alternatively, download the source zip as described above.
Next, in the command window, cd to the folder you extracted the source files to.
From here cd to 'h5serv-master/server' (or just server if you used git).
Run: python app.pyYou should see the output: 'Starting event loop on port: 5000'.
You may then see a security alert: 'Windows Firewall has blocked some features of thisprogram'. Click 'Allow access' with the default option (Private network access).
At this point the server is running, waiting on any requests being sent to port 5000.Go to the 'verification' section below to try out the service.
Installing on Linux/Mac OS X¶
Anaconda from Continuum Analytics can be used to easily manage the package dependenciesneeded for HDF Server.
In a browser go to: http://continuum.io/downloads and click the 'Mac OS X 64-bitPython 2.7 Graphical Installer' button for Mac OS X or: 'Linux 64-bit Python 2.7'.
Install Anaconda using the default options.
Once Anaconda is installed, open a new shell and run the following on the command line:
condacreate-nh5servpython=2.7h5pytornadorequestspytz
Answer ‘y' to the prompt, and the packages will be fetched.
Install the watchdog package (this is currently not available through Anaconda):pipinstallwatchdog
Deck the balls mac os. In the same shell, run: sourceactivateh5serv
Download the hdf5-json project: gitclonehttps://github.com/HDFGroup/hdf5-json.git .Alternatively, in a browser go to: https://github.com/HDFGroup/hdf5-json and click the'Download ZIP' button (right side of page). Download the zip file and extract tothe destination directory of your choice.
Next, cd to the hdf5-json folder and run: pythonsetup.pyinstall.
Download the h5serv project: gitclonehttps://github.com/HDFGroup/h5serv.git .Alternatively, download the source zip as described above.
Next, cd to the h5serv folder.
From here cd to 'server' (or 'h5serv-master/server' if you extracted from ZIP file).
Run: pythonapp.pyYou should see the output: 'Starting event loop on port: 5000'.
At this point the server is running, waiting on any requests being sent to port 5000.Go to the 'verification' section below to try out the service.
Verification¶
To verify that h5serv was installed correctly, you can run the test suite includedwith the installation.
Open a new shell (on Windows, run 'Annaconda Command Prompt' from the start menu).
In this shell, run the following commands:
- source activate h5serv (just: activate h5serv on Windows)
- cd
- cd test
- python testall.py
All tests should report OK.
Server Configuration¶
The file h5serv/server/config.py provides several configuration options that can beused to customize h5serv. Each of the options can be changed by:
- Changing the value in the config.py file and re-starting the service.
- Passing a command line option to app.py on startup. E.g.
pythonapp.py--port=7253 - Setting an environment variable with the option name in upper case. E.g.
exportPORT=5000;pythonapp.py
The config options are:
port¶
The port that h5serv will listen on. Change this if 5000 conflicts with another service.
Default: 5000
debug¶
If True the server will report debug info (e.g. a stack trace) to the requester onerror. If False, just the status code and message will be reported.
Default: True
datapath¶
A path indicating the directory where HDF5 files will be be stored.
Note: Any HDF5 file content that you put in this directory will be exposed via theserver REST api (unless the domain's ACL is configured to prevent public access, see:./AclOps).
Default: ./data/
public_dir¶
A list of directories under datapath which will be visible to any autenticated user'srequest.
Default: ['public','test']
domain¶
The base DNS path for domain access (see comment to hdf5_ext config option).
Default. hdfgroup.org
hdf5_ext¶
The extension to assume for HDF5 files. The REST requests don't assume an extension, soa request such as:
Translates to: 'Get the file tall.h5 in the directory given by datapath'.
Default: .h5
toc_name¶
Name of the auto-generated HDF5 that provides a 'Table Of Contents' list of all HDF5files in the datapath directory and sub-directories.
Default: .toc.h5
home_dir¶
A directory under data_path that will be the parent directory of user home directores.For example if datapath is ./data, home_dir is home, the authenticated requestof GET/ for userid knuth would return a list of files in the directory:./data/home/knuth.
Default: home
ssl_port¶
The SSL port the server will listen on for HTTPS requests.
Default: 6050
ssl_key¶
The SSL key.
default:
password_uri¶
Resource path to be used for user authentication.Currently two methods are supported:
HDF5 Password file: An HDF5 that contains userids and (encrypted) passwords.See: Admin Tools. In this case the password_uri config is a pathto the password file.
MongoDB: A MongoDB database that contains a 'users' collection of userids andpasswords. In this case the password_uri would be of the form:mongodb://: where is the IPaddress of the host running the mongo database and is the port ofthe mongo database (typically 27017).
default: ./util/admin/passwd.h5
mongo_dbname¶
Mongo database named used for MongoDB-based authentication as described above.
default: hdfdevtest
static_url¶
URI path that will be used to map any static HTML content to be displayed by the server.
default: /views/(.*)
static_path¶
Overrun Debug 0 1 Mac Os Catalina
File path for files (i.e. regular HTML files) to be hosted statically.
default: ./static
cors_domain¶
Domains to allow for CORS (cross-origin resource sharing). Use * to allowany domain, None to disallow.
default: *
log_file¶
File path for server log files. Set to None to have logout go to standard out.
log_level¶
Verbosity level for logging. One of: ERROR,WARNING,INFO,DEBUG,NOTSET.
default: INFO
background_timeout¶
Time interval in milliseconds to check for updates in the datapath folder (e.g. a filethat is added through some external process). Set to 0 to disable background processsing.
default: 1000
Data files¶
Copy any HDF5 files you would like exposed by the service to the datapath directory(h5serv/data). If you do not wish to have the files updatable by the service make thefiles read-only.
Overrun Debug 0 1 Mac Os Update
On the first request to the service, a Table of Contents (TOC) file will be generated whichwill contain links to all HDF5 files in the data folder (and sub-folders).
Overrun Debug 0 1 Mac Os Download
Note: Do not modify files once they have been placed in the datapath directory. h5servinventories new files on first access, but won't see some changes (e.g. new group is created)made to the file outside the REST api.
*Note: HDF5 that are newly created (copied into) the datapath directory will be 'noticed'by the service and added into the TOC.
