de.NBI Cloud Berlin¶
Welcome to the de.NBI Cloud site Berlin. In the following guide we want to give you a quick introduction how to use our cloud site.
Please note, you are responsible for everything that happens with the virtual machines (VMs) you deploy! We as resource provider are not liable for anything and do not give any guarantees. Keep this in mind and try to make your setup as reproducible as possible from the beginning.
How to get in contact with us¶
In case you have questions or want to give us any kind of feedback, please contact us via firstname.lastname@example.org.
Access to the de.NBI Cloud Berlin¶
The OpenStack dashboard gives you all information about your project, e.g. your available resources, your virtual machines etc. The dashboard is available here. To access your project, use Elixir as authentication provider. After authentication you will be redirected to the OpenStack dashboard.
Deploy your VMs¶
The networks are already preconfigured, so you can directly start and deploy your VMs. Therefore go to Project - Compute - Instances and choose the button "Launch Instance" on the right side. Now you have to provide information at least in the categories Details, Source and Flavor.
- In the field "Instance Name", assign a name to your VM
- Choose "Image" as "Boot Source"
- If you choose "Create New Volume" No, the vm volume will be created on hypervisor and be limited to 20GB.
- If you choose "Create New Volume" Yes, your volume will be placed on our separate storage system and can be bigger than 20GB. Please to not create volumes bigger than 250 GB! (You can use a share instead)
- Choose an appropriate image from the list (e.g. CentOS)
- Choose a Flavor from the list that fits your resource requests (e.g de.NBI default provides 2 cores with 4GB of RAM)
- For new projects (09/2021) it is recommended to choose "yourProject-network-2". You will be able to use floating ips within subnet of public2:
172.17.0.10 - 172.17.7.250
- Older projects (prior 09/2021) can still use their existing "your-Project-network". You will be able to use floating ips within subnet of public:
172.16.102.200 - 172.16.103.220
Now hit the button "Launch Instance". The VM will be deployed and accessible in a few seconds. To connect to your VM you need to assign a floating ip address to the machine. Therefore click on the arrow on the right side of the spawning VM, choose "Associate Floating IP" and use one available floating ip addresses from the drop-down menu.
Hint: To connect to one of your VMs without a floating ip address you have to assign at least one floating ip address to another of your machines. As soon as you are connected to this machine you are inside of your project network and can connect to VMs without any floating ip address.
Connect to your VMs with ssh¶
None of your VMs will be directly visible and accessible from the internet. To connect to one of your VMs, you have to use our jumphost server denbi-jumphost-01.bihealth.org with your elixir login name (not your elixir id!).
Your public ssh must be added at the de.NBI portal: https://cloud.denbi.de/portal/ (or at the elixir portal) to able to connect to our jumphost. and make sure that you import a public ssh-key into your OpenStack project (Project - Compute - Key Pairs - Import Key Pair) so that you can access your VMs later on.
Please take care, as for now, that our images are shipped with the standard users for the respective Linux distribution. Here you can see a list of standard users for some common distributions:
- CentOS: centos
- Ubuntu: ubuntu
Host denbi-jumphost-01.bihealth.org HostName denbi-jumphost-01.bihealth.org User ElixirLogin IdentityFile PATH_TO_KEY Host NAME_OF_VM # first vm HostName 172.16.XXX.XXX IdentityFile PATH_TO_KEY User ubuntu / centos ProxyCommand C:\Windows\System32\OpenSSH\ssh.exe denbi-jumphost-01.bihealth.org -W %h:%p
If you want to connect from a Windows-based system you can use Putty (http://www.putty.org/) to connect to the jumphost and your virtual machines.
In the field Host Name you have to enter the name of our jumphost:
Under Connection - Data you can choose the username for the auto-login. Please use your elixir username here.
In the section Connection - SSH - Auth you can provide your SSH-key. Please make sure that you also check the option Allow agent forwarding so that you can connect to your VM.
When you connect to our jumphost for the first time you may get a warning about accepting the servers host key. Please confirm with yes.
Host denbi-jumphost-01.bihealth.org HostName denbi-jumphost-01.bihealth.org User ElixirLogin IdentityFile PATH_TO_KEY ServerAliveInterval 120 Host NAME_OF_VM HostName 172.16.XXX.XXX IdentityFile PATH_TO_KEY User ubuntu / centos ProxyJump denbi-jumphost-01.bihealth.org
Connection with SSH-Agent¶
Manually jump from your client to jumphost and from there further to your vm with ssh-agent-forwarding
eval `ssh-agent` // eval $(ssh-agent)
Add ssh private key:
bash ssh-add .ssh/id_rsa
bash ssh-add -l
Connect at first from your client to jumphost:
ssh -A yourElixirLogin@denbi-jumphost-01.bihealth.org
And from the jumphost you can connect further to the floating ip of your vm
Setting up a SOCKS proxy¶
In some cases it would also make sense to configure a permanent SOCKS proxy to communicate with your VMs behind the jumphost, e.g. when using web applications etc. As long as you have an open SOCKS connection to the jumphost you can directly connect to your VMs from a different console. In the following example socat is used but also netcat (nc) works in a similar way. Add the following lines to your local ~/.ssh/config:
# Access to the de.NBI jumphost Host denbi-jumphost-01.bihealth.org # Use your Elixir login name User ElixirLogin # Use your ssh-key file IdentityFile YOUR-SSH-KEY-FILE # Open a SOCKS proxy locally to tunnel traffic into the cloud environment DynamicForward localhost:7777 # Forward locally managed keys to the VMs which are behind the jumphosts ForwardAgent yes # Send a keep-alive packet to prevent the connection from beeing terminated ServerAliveInterval 120 # Access to de.NBI cloud floating IP networks via SOCKS Proxy Host 172.16.102.* 172.16.103.* # Tunnel all requests through dynamic SOCKS proxy ProxyCommand /usr/bin/socat - socks4a:localhost:%h:%p,socksport=7777 or socks5h:localhost:%h %p,socksport=7777 # Use your ssh-key file IdentityFile YOUR-SSH-KEY-FILE # Forward locally managed keys ForwardAgent yes # Send a keep-alive packet to prevent the connection from beeing terminated ServerAliveInterval 120
If you need more disk space than the initial image provides (20GB), one way is to create a volume and attach it to your VM. Please keep in mind that a volume can only be attached to one VM at the same time. The advantage of a volume is that it will be available also after you deleted your VM. So you can use it to store data temporally. Please to not create volumes bigger than 250GB
To create a volume choose Project - Compute - Volumes followed by Create Volume on the right side. Now assign a name to your volume and set the size according to your needs. After the successful creation of the volume you have to attach it to your VM. Choose the arrow on the right side of the created volume and select Attach Volume. In the new window you have to choose your VM from the drop-down menu under Attach to Instance.
On your VM you now have to make a filesystem on the device so that you can mount it to your machine. Use e.g. mkfs to make an ext4 filesystem:
sudo mkfs.ext4 /dev/vdb
After creating the filesystem you can mount the filesystem to your VM and start using it:
sudo mount /dev/vdb /mnt
Create a NFS share¶
In case you need a NFS share to store big amounts of data and share it within your project, you can use OpenStack to create and manage the share.
To create a NFS share choose the section Shares and click on Create Share. In the popup you have to provide the following information:
Share Name: - Provide a share name.
Share Protocol: - Please use the preselected "NFS" as protocol.
Size (GiB): - Provide the size of the share. Info: You have an overall quota for NFS shares on your project. Please make sure that you set the size below the project quota.
Share Type: - Please select "default" (or dmz, if you want to use a share from a server within the dmz)
Availability Zone: - Please select "nova".
Manage access rules for your NFS share¶
After the creation of a NFS share, the share will not be accessible by anyone . To grant your VMs access to the share you have to configure the access rules.
Important: Please make sure to keep the access rule list of your NFS share up to date, so that only your VMs can access the share.
To manage the access rules click on the arrow on the right side of your newly created NFS share and choose Manage rules. Now you have to choose Add rule. In the popup you have to provide the following information:
Access Type: - Select ip to allow a certain VM access to the share.
Access Level: - Choose read-write or read-only appropriate to your needs. In some cases it may make sense that specific VMs just get read-only permissions.
Access to - Please fill in the ip address of your VM you want to grant access to the NFS share.
Access your NFS share¶
In order to use your created NFS share you have to mount it to your VMs. Click on the created share in the Shares section of the OpenStack dashboard to get information about the complete mount path. Under the Export locations section, please choose the Path e.g.:
You can mount the share with the following command:
sudo mount -o vers=4.0 manila-prod.isi.denbi.bihealth.org:/ifs/denbi/prod/share-YOUR_UIID /mnt/
Alternatively you can add the mount path to the "/etc/fstab". Make sure that you use NFS version 4.0. In order to use NFS version 4.0, you might need to set the host's DNS domain name:
cat /etc/idmapd.conf [General] #Verbosity = 0 # The following should be set to the local NFSv4 domain name # The default is the host's DNS domain name. Domain = denbi.bihealth.org
Please make sure that your user (depending on the used distribution: centos, debian, ubuntu) is the owner of the NFS share. Therefore run the following command to set the user as owner of the NFS share:
sudo chown centos:centos /mnt/
Hint This example is for a Centos based image.
Using the OpenStack API¶
First, you will need to request a password to use the OpenStack API, therefore write a mail to the support team at email@example.com. Second, the API is not directly accessible from the outside, so the only way to access the API from a local machine is through the jumphost. So make sure you've configured your SOCKS proxy as described before. In addition you will need to configure your environment to use the SOCKS proxy for the API requests. Therefore set your environment variables for the http/https proxy:
export http_proxy=socks5h://localhost:7777 export https_proxy=socks5h://localhost:7777 export no_proxy=localhost,127.0.0.1,::1
Now, if you have an active SOCKS connection to the jumphost, you should be able to use the OpenStack API from your local machine.
Adding multiple SSH-Keys¶
To access your VM you have to provide a public ssh-key. In the deployment step of your VM you can choose which public ssh-key you want to use for your VM in the section Key Pair.
In case you want to directly deploy a VM and give access to more than one user you can use the section Configuration - Customization Script in the deployment part. Here you have to list the full public keys in the following format:
#cloud-config ssh_authorized_keys: - Full public ssh-key of User-1 - Full public ssh-key of User-2
After the successful deployment of the VM, user 1 and user 2 will have access to the VM.
Upload your own Linux images¶
If you need an extra Linux image we do not provide, you also can upload your own images via Project - Compute - Images. Select Create Image and choose a name and the path for the image and also make sure that you choose the correct format (typically qcow2). If there are special requirements for your image, you can specify the minimum disk size and also the minimum amount of RAM. After the successful upload only the members of your project can use the image.