So that each project could use one SSD disk.
I am assuming that your question pertains to SSD devices for particle caching. This capability is not currently implemented. Thanks for the suggestion @dingwei.
Did you consider combining multiple SSDs into a single, larger volume that would be dynamically shared for particle caching by all projects?
Please let us know whether your question concerned storage for project directories instead of particle caching.
[Updated: clarified storage purpose]
Do you mean the main project directory, or a scratch drive, or the database?
As long as the user who is running CryoSPARC has read/write permissions to a disk, there is nothing stopping you saving a project there (if you know the mount point). But if it’s a removable drive (e.g.: SSD in USB dock) CryoSPARC gets unhappy unless you detach the project before removing it, for obvious reasons…
Thank you for your patient reply. I have tried to combine multiple SSDs into one big disk for different projects. However, the interference between them is significant. When one project is caching particles to the disk, the jobs in other projects become extremely slow. What’s even more concerning is that the particles cached by one project can be deleted by another project without any warning. That is why I hope to run each project on a specific SSD to avoid these issues.
The cache system is designed so that, when a job is running, not-in-use (but see caveat below) cached particles are deleted at the earlier of
- the configurable variable
$CRYOSPARC_SSD_CACHE_LIFETIME_DAYS(guide) - there is not enough room on the cache volume to accommodate the current job’s cache request
As long as cache storage is not shared between worker nodes, a running job’s particles should never be deleted.