CryoSPARC Integration

Summary

CryoSPARC integration issues cover installation, upgrades, runtime startup failures, scratch-storage configuration, GPU/driver compatibility, and post-RMA reconfiguration on Exxact systems. Most cases are software-stack or environment mismatches rather than hardware defects, though a few begin with hardware-like symptoms.

Frequency

• 35 tickets in this evidence set mention CryoSPARC-specific setup, compatibility, access, or runtime integration problems.

Common Causes

1. Hostname, node-role, or local configuration mismatch. Common symptoms include master-node errors, broken hostname checks, stale GPU counts, or services pointing at old settings. (#15173, #9859, #9414, #39626, #16625, and 5+ more)
2. Driver, CUDA, or OS compatibility drift. CryoSPARC upgrades or reboots often exposed NVIDIA/CUDA/DCGM mismatches, unsupported Rocky/CentOS combinations, or kernel-driver issues. (#16198, #23329, #26473, #28267, #37988)
3. Scratch, data, or storage-layout confusion. Customers hit missing scratch disks, wrong cache mounts, full partitions, or permission problems that CryoSPARC surfaced first. (#13972, #5083, #5210, #5627, #9295)
4. Install or reinstall workflow gaps. Several tickets center on fresh installs, reinstall scripts, directory guidance, or uncertainty about what Exxact does versus what the customer must configure. (#16273, #17299, #18254, #19468, #23632, and 5+ more)
5. CryoSPARC-specific application/runtime faults. Stale locks, expired licensing, OOM behavior, or portal/login issues appeared even when the underlying hardware was fine. (#12241, #5270, #20213, #39681, #39621)

Diagnostic Steps

1. Confirm the failure is inside CryoSPARC, not hardware first. Check whether GPUs appear in nvidia-smi, whether storage is mounted, and whether the OS itself is stable. (#39626, #39621, #16198, #5083)
2. Validate CryoSPARC environment and config. Check hostname/master-node settings, service status, shell config, install path, and any recent updates or reinstall attempts. (#12241, #15173, #9859, #29890, #9414)
3. Check software-stack compatibility. Review CUDA, NVIDIA driver, DCGM, OS version, and whether the CryoSPARC release is supported on that platform. (#16198, #23329, #26473, #28267, #16273)
4. Inspect storage assumptions. Verify scratch path, free space, mount points, permissions, and whether CryoSPARC is targeting the intended disk. (#13972, #5083, #5210, #5627, #9295)

Solutions

1. Correct local configuration or reinstall CryoSPARC cleanly. This solved hostname, master-node, shell, and service-state issues when the underlying hardware was fine. (#12241, #15173, #9414, #9859, #29890, and 5+ more)
2. Repair the driver/CUDA stack to match CryoSPARC requirements. Driver recovery, CUDA updates, or OS-version changes were the common path for GPU-runtime failures. (#16198, #16273, #23329, #26473, #37988)
3. Fix scratch/data path layout. Working fixes included using the correct scratch path, symlinking to the real scratch location, adding storage later, or cleaning the full partition CryoSPARC depended on. (#13972, #5083, #5210, #9295, #5627)
4. Use guided installation artifacts when setup is the blocker. Successful cases often ended with a validated install script, official docs, or a known-good configuration path rather than ad hoc commands. (#16273, #17299, #18254, #19468, #23632)
5. Reconfigure CryoSPARC after hardware or RMA changes. Some follow-ups were resolved only after updating CryoSPARC to match the current GPU count or repaired system state. (#39621, #39626, #16625, #22872)

Edge Cases

• Hardware symptom that turned out to be CryoSPARC config. GPU-count or crash reports sometimes resolved once CryoSPARC was reconfigured rather than replacing parts. (#39621, #39626)
• CryoSPARC issue layered on top of a separate boot or RMA problem. A few mixed-scope tickets started as system-return or boot problems and only later narrowed to CryoSPARC integration. (#16625, #22872, #39621)
• Customer self-resolution or partial closure. Some tickets closed after the customer fixed the lock state, updated locally, or accepted advisory guidance without a deeply documented final test. (#12241, #26473, #30594 is related style but out of set, #22722)
• Process-quality risks. Several tickets note public credential sharing, unclear scope boundaries, or long asynchronous loops that made software resolution harder than it needed to be. (#20489, #22872, #8913, #28267)

Related Issues

• Software Installation
• Firmware Driver Compatibility
• BIOS BMC Issues
• Defective Storage Drives
• Credential Recovery

Referenced by

• Software Installation — co-occurs with this issue (×11)
• Matt — handled tickets on this issue (×5)
• RTX A5000 — product affected by this issue (×1)
• Garry Gayles — handled tickets on this issue (×3)
• Andrew Rodriguez — handled tickets on this issue (×8)
• Nam Luong — handled tickets on this issue (×7)
• RTX 3090 — product affected by this issue (×1)
• Duc Bui — handled tickets on this issue (×2)
• Defective Storage Drives — related issue (×2)
• RMA Workflow — co-occurs with this issue (×3)