CryoSPARC jobs fail after temporary network disconnection on a single workstation

Install, Configure and Update
1

Dear CryoSPARC Support Team,

I am running CryoSPARC v5.0.6 on a single workstation where both the master and worker are installed on the same machine. The workstation is protected by a UPS, so it does not lose power during interruptions.

However, when the network connection is temporarily lost (for example, due to a switch/router power interruption), all running CryoSPARC jobs eventually fail. CryoSPARC services remain running, and increasing the timeout settings only delays the failure:

CRYOSPARC_HEARTBEAT_SECONDS=3600
CRYOSPARC_CLIENT_TIMEOUT=3600

The jobs continue running during the network outage but fail after the heartbeat timeout with a lost connection error.

During troubleshooting, I noticed that some internal CryoSPARC connections are using the workstation’s network IP address rather than only the loopback interface (127.0.0.1), even though both master and worker are on the same machine.

Would it be possible and supported to configure a single-workstation installation to use:

CRYOSPARC_MASTER_HOSTNAME="127.0.0.1"

or

CRYOSPARC_MASTER_HOSTNAME="localhost"

to avoid dependency on the physical network interface?

Before making this change, I would like to confirm:

  1. Is this configuration supported for a single-workstation CryoSPARC installation?

  2. Could changing CRYOSPARC_MASTER_HOSTNAME affect existing projects, databases, or stored data?

  3. Is there a recommended procedure to make this change safely?

Thank you very much for your help.

Best regards,

Daniel

2

@PhoX Thanks for your question. Another potentially helpful setting might be an increased database connection timeout (in milliseconds), which is 20 seconds by default:

export CRYOSPARC_DB_CONNECTION_TIMEOUT_MS=20000

is possible (with some additional changes), but depending on your computer’s other uses, you might instead consider

  1. setting CRYOSPARC_MASTER_HOSTNAME to your computer’s full hostname. Changes in cryosparc_master/config.sh require a CryoSPARC restart to become effective.
  2. adding a record to /etc/hosts that resolves your computer’s full hostname to a loopback address, like 127.0.1.1, as discussed here. This approach may be problematic under some circumstances (google).