Data Practices
The majority of our data is kept on the network file system on Graham, as this is regularly backed up. Simlarly, our code and manuscripts are version-controlled in appropriate locations (see below for options).
Store your work somewhere where it is backed up and if possible, version controlled
Data Storage
The Digital Alliance of Canada has different tiers of storage: /home
, /project
, and /scratch
/home
- store copies of your code and other small files/project
- where project data will be coming from; use this folder for finalized data
The /project
filesystem has a STRICT limit on the total number of files storage size, so you should zip/tar your files that reside there to ensure the average file size is at least 1GB
/scratch
- where active work should be done
Files on the /scratch
filesystem expire after 60 days if they have not been modified. You will be sent 2 e-mails prior to the deletion of these fies. Move your data to /project
once it has been finalized!
Code
Code should be stored in a git repository (preferably on GitHub, using either a public or private repository). Clone repositories can then be in your home directory, local machine, or scratch space. If you make a mistake, you can revert to a previous commit (e.g. saved version).
Store code, scripts, and markup documents here. Changes made should be committed and pushed to GitHub (e.g. time-stamped and associated with user).
Manuscripts
Western provides a OneDrive account that can be used to store papers you are reading or writing, as well as any other reference material. Alternatively, you can store these on Google Drive. These tools make it easy to share and collaborate on documents.