our deadline server has enough room but not so much that we can afford to keep jobs lying around, even archived. but! the historical data in the database is very valuable to us in terms of making sure we’ve improving efficiency over time. is there a way to remove the actual job files (max files, supplementary textures) from the repository but leave the old job data (name, aux file size, submission times and all that stuff that lives in the db)?
sidenote: my question above is predicated on the assumption is that the ‘delete job’ command deletes both the job files from the repository /and/ the job data from the database. please let me know if that’s not correct and the data lives on in the database.
boy, i mangled the topic. that should read: “keeping old render data without keeping jobs or their archives.”
Hello,
I am thinking that perhaps an event plugin to delete scene files on completion would work here, maintaining the data about the job, without the overhead of the scene files. Or perhaps moving them off server to ensure you don’t completely lose the scenes. Just a thought, if you have a coder on staff. If not, we can try to see what’s possible on our end.
we opted to run a scheduled task that concatenated the current db to a duplicate db with aggregate data (not deleting old entries), which let us not have to keep hundreds of old jobs active. basically a mongodump and mongorestore sequence.
Not a bad idea. I’d do the programming approach with events to store the job data after the job was deleted, but there’s more than one way to solve this sort of problem. 