I have a question about setting up the cryosparc master-worker system on ubuntu. We have multiple GPU workstations. Currently, each GPU workstation is an independent master and they have different IP addresses to login cryosparc. But they all write into the same cryosparc database. I can attach projects from one workstation to another. I think we meet all three criteria for master-worker architecture (unified file system, passwordless ssh, TCP ports available). I wonder if there is an easy way to transition from the current architecture into master-worker system without damaging the current cryosparc database. If so, what would be the best way to do this?
Any help would be appreciated.
Thanks in advance.
Chen
Do all the CryoSPARC masters/instances write to the same (mongo) database directory that is shared between the workstations? Sharing a database between CryoSPARC masters would result in CryoSPARC malfunction and/or data corruption and/or undefined behavior of the software.
Thank you so much for your reply. Everyone’s cryosparc project folder is in the same folder on our file system. I believe they all write to the same mongo database directory. Otherwise how could I attach the project from one to another without exporting and importing jobs? Maybe that’s also why we always run into errors when we use SSD to cache data. We have to disable caching data to SSD for every job we run. Can you provide more info on the disadvantages of sharing a database between CryoSPARC masters for our reference?