Unfortunately, it looks like you may have kept the 72207 job private, which means I can’t look at your job in particular to help debug things.
But generally speaking, the thing to do for failed ROSIE jobs is to take a look at the `[Full set of decoy structures created on this run (FS view)] ` link on the page, specifically the `logs` subdirectory there. There may be more information there about what might be going on.
Looking at some of the failed jobs in the VIP queue, one possibility for the failure is the runtime. To keep ROSIE from being bogged down with really big jobs, the VIP jobs are limited to 24 CPU hours of runtime. If you’re getting greater than 24 under “CPU hours used”, you may need to adjust how you’re running things (e.g. cutting the submitted protein down just to a single relevant domain, reducing the number of cycles or tries, or adding more excluded residues).
.
