Setting up Volumes on Jetstream
May 23 2018
Why use volumes
Volumes are like remote hard drives to attach to virtual machines (VM) to save input data, outputs, and important results.
Few scenarios where having this option is awesome! and spares you a lot of time
- I have run into this problem (several times) where I spin up a large VM which provides 30GB memory and 60GB disk space. This VM provides enough memory for running my analysis, but I have close to 60GB of data or more to run through the analysis, leaving no space to save my output files. Instead of now having to kill this VM and spin up a larger VM, there is another option. Spin up a volume instead, it’s like a hard drive, where you can save your input files and output files and use the VM to run the analysis.
- Using Jetstream VM to share data between collaborators. No analysis needs to be done, just sharing data. Spin up a small VM, attach a volume with data to be shared. Give collaborators access to the VM.
The advantage to using a volume is that
- Setting up volumes does not use up your allocation (or allocated SU’s).
- Each user can get 10 volumes up to 500GB total storage and can request for more space by sending the Jetstream team a mail, without having to alter your allocation request.
- You can spin up a volume after starting your analysis on the VM and attach it then.
Note: A volume can be attached to only ONE VM at a time.
Spin up a volume and attach to VM
Follow these steps
- Make sure the VM to whom you would like to attach the volume to, is active
- Spin up a volume (above figure),
MAKE SURE that the volume is from the same provider as the VM. In the image below the purple boxes show where the service provider information. There are two providers – IU and TACC, you cannot connect a volume at IU with VM provided by TACC, and vice-versa
Also, notice that the volume status says ”Unattached” in the above diagram
Click on volume, and select “Attach” on the right corner of the screen
There will be a pop-up screen with a list of all the active VM’s in your account and you can select one.4. Now wait for the status to change to “Attached”
Accessing the Volume and moving data here
- Using the Wed Desktop, open “File Manager”, then select “Filesystems”, there should be a “ vol_*”, as shown below
Note: The reason I use * instead of ' b' in vol_b above is that sometimes, the volumes can be named vol_c as well. In this case, the volume named "vol_b"
In WebShell or ssh into the VM
If you are not sure what your volume name is, the type
ls #lists out the files, and there should be directory names vol_*, as shown in the below image
- You can move data here to this volume from your VM using commands like mv or cp bash commands.
Data transfer to the volume directly
- Make sure volume is attached to the VM.
- Use any data transfer commands like rsync, scp, ftp to move data from cluster/laptop to volume directly. For more information on these commands goto here
- There is always Globus, if you are new to Globus – follow this blog post first
Then set up an endpoint on the VM. To do that – follow this blog post but with a small tweak in Step 8. Following the blog without the small change given below NOT give you access to the volume attached to the VM.
Step 8 – with the small tweak to get the volume to show up on Globus, follow along or follow globus documentation . As a note the words in red are commands that can be copied, the words in purple have to be replaced with your informationNow setup endpoint on globusonline.org for your instance
endpoint-add -n name_of_the_endpoint -- gc
This command will give you an Endpoint ID and Setup ID for the instance, which you WILL NEED in the next set of commands. Copy paste them into a text editor if needed. Sign out of the ssh with "exit", then in the globus install location:
./globusconnect -setup Setup_ID
nano ~/.globusonline/ lta/config-paths #opens a file editor nano, you can use your text editor of choice
Type this line in the text editor, make the change to include the correct volume name.
Save the file in nano – Ctrl and O (at the same time)
Exit text editor in nano – Ctrl and X (together)
Start globus service
./globusconnect -start &
./globusconnectpersonal -start &
Now on the globus application, the Volume should show up as a directory
Deleting VM and Volume
First, make sure to back up data (input files, outputs) on the VM/volume to your computer/data archive. Once you delete the volume and VM, data cannot be recovered.
- Next, Detach the Volume from VM
- Then delete VM and volume
If you have any questions, send us an email at firstname.lastname@example.org