Because I don’t know more about the setup, use(r) patterns and workloads, I can only provide general pointers to
- network issues
- computer or network-wide firewalls
- network stability
- network services (like DNS)
- workload
- computer may be overwhelmed by demands on memory, processing, storage
If any of these issues results in an “unclean” disruption of cryoSPARC, additional issues may arise, such as “zombie” processes or stale socket file(s). The latter problem would manifest itself in the presence of /tmp/mongodb-*.sock and/or /tmp/cryosparc-supervisor-*.sock files owned by the linux user running the cryoSPARC instance even though the cryoSPARC instance is down and the supervisord, mongod processes have exited.