EO DATA resources might be mounted to VM through s3 interface using s3fs command.

1. Perform operation as a root

# Perform all commands as root
sudo -i

2. Install build and runtime dependencies for S3FS

For  Ubuntu/Debian

# Install git
apt-get update
apt-get --yes install git
# Install required tools and libraries
apt-get --yes install automake autotools-dev fuse g++ git libcurl4-openssl-dev libfuse-dev libssl-dev libxml2-dev make pkg-config

For CentOS/RedHAT

yum update
yum install git
yum install automake fuse fuse-devel gcc-c++ git libcurl-devel libxml2-devel make openssl-devel

3. Build s3fs

git clone https://github.com/CloudFerro/s3fs-fuse.git
cd s3fs-fuse/
./autogen.sh
./configure
make
# Install s3fs
make install

4. Mount /eodata using s3fs

# Prepare configuration for EO Data
echo ACCESS:DATA >  ~/.passwd-s3fs
chmod 600 ~/.passwd-s3fs

# Comment out original fstab entry that mounted /eodata over NFS
sed -i 's/nfs.eodata/#nfs.eodata/g' /etc/fstab

# Add new entry to mount /eodata over s3fs
echo s3fs#DIAS /eodata fuse passwd_file=/root/.passwd-s3fs,_netdev,allow_other,use_path_request_style,uid=0,umask=0222,mp_umask=0222,gid=0,url=http://data.cloudferro.com/ 0 0 >> /etc/fstab

# Modify /etc/updatedb.conf to exclude s3fs mounts from indexing
sed -i 's/PRUNEPATHS="/\PRUNEPATHS="\/eodata /g; s/PRUNEFS="/PRUNEFS="s3fs /g' /etc/updatedb.conf

5. Mount /eodata using s3fs

# Remount /eodata
umount -lf /eodata
mount /eodata

The whole mount process on new VMs is automatic so if you don't want it to be done automatically, just remove the option noauto form /etc/fstab file.

s3fs#DIAS   /eodata fuse    noauto,_netdev,allow_other,use_path_request_style,uid=0,umask=0222,mp_umask=0222,mp_umask=0222,gid=0,url=http://data.cloudferro.com,use_cache=1,max_stat_cache_size=60000,list_object_max_keys=10000,comment=cloudconfig    0   0

By default, S3FS Cache is enabled in CREODIAS preconfigured Linux images. To examine its directory location you might take a look at the configuration in /etc/fstab.

#s3fs#DIAS /eodata fuse passwd_file=/root/.passwd-s3fs,_netdev,allow_other,use_path_request_style,uid=0,umask=0222,mp_umask=0222,mp_umask=0222,gid=0,stat_cache_expire=20,url=http://data.cloudferro.com,use_cache=1,max_stat_cache_size=60000,list_object_max_keys=10000 0 0

Use_cache parameter points on the folder “1” (it is located at the beginning of Filesystem Hierarchy /)
From the manual pages:

-o use_cache (default="" which means disabled)
local folder to use for local file cache.

Directory “1” is being filled up during processing or downloading products from EOData repository to your local storage directly.

If you encounter problems with displaying the folders and/or files inside "/eodata" (e.g. "no such file or directory") please empty the "/1" folder.

To avoid the situation when you cannot remount your resources because the cache directory is occupying a whole left space on the disk, you might try to set a parameter ensure_diskfree.

Common prompt messages when you cannot remount an /eodata:

clnt_create: RPC: Timed out
s3fs: There is no enough disk space for used as cache(or temporary) directory by s3fs.

From s3 manual pages, you can find out about ensure_diskfree parameter.

-o ensure_diskfree(default 0)
	sets MB to ensure disk free space. This option means the threshold of free space size on disk which is used for the cache file by s3fs. s3fs makes file for downloading, and uploading and caching files. If the disk free space is smaller than this value, s3fs do not use diskspace as possible in exchange for the performance.

If you are using services that are billed in Pay-Per-Use mode, you can experience the situation, when your credits would be running low.

In such case, you will receive email notification from our billing system when your account balance will be lower than 50 credits.

If you run out of credits, you will get another email.
If your account overdraft will fall below -100  credits, all your instances will be paused.
If you do not buy additional credits, your virtual resources will be deleted after 30 days.

You can always buy new credits through our site here
If you have any problems or questions please contact our support

To create SSH key pair (private and public) in Windows you need to install puttygen.exe application.
You can download it from http://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html.
After installation please run puttygen.exe.
 

Click "Generate" button and move the mouse cursor randomly over the blank area.

You will obtain the following screen.
 

Mark and copy the text from the field under the title "Public key for pasting into OpenSSH authorized_keys file".
Open the notepad and paste the whole text into new file.

Save the file in the location of your choice (eg. C:\temp) and give the file a name (eg. key-01.pub).

ATTENTION: the above method works well with Openstack.
If instead of copy-and-paste you save the public key by clicking "Save public key" in puttygen.exe, the key will not be accepted by Openstack (it has different format from that in the field under the title "Public key for pasting into OpenSSH authorized_keys file").

Login to https://creodias.eu and go to Compute  -> Key Pairs

Click "Import Key Pair".

Give the new Key Pair the name (eg. key-01).
Paste the public key (obtained with puttygen.exe and saved in separate file) into Public Key field.

Click "Import Key Pair" button.
If you go to Compute -> Access & Security, you will see the new Key Pair on the list.

Come back to puttygen.exe application and click "Save private key" button.
Save the key in the location of your choice (eg. C:\temp) and give the key name (eg. key-01-priv).
The private key should be saved with the extension .ppk (eg. key-01-priv.ppk).

Using the private key and the puTTY.exe application you can login to the VM.

If you would like to learn more about PuTTYgen, learn about the installation or usage guide, we invite you to visit the website www.puttygen.com