Subset Particles by Statistic job no longer fails when manual thresholds are specified in the job builder (previously only included in v4.7.1-cuda12+250811 patch)
Installation
See installation instructions here (CryoSPARC v4.7.1):
This patch fails to install on my master, while the worker was patched correctly the first time. I’ve tried downloading the file first and then installing it, and I’ve tried the --force flag, but still no success. The output doesn’t indicate any problem:
$ cryosparcm patch --force
A cryoSPARC patch is available
Current Version: v4.7.1
Current Patch: None
New Patch: 250814
Released On: 2025-08-14 20:14:16
Requires Restart: Yes
Patch Notes:
- Fixed: Ab-Initio Reconstruction no longer fails with a NaN error when one or more densities become completely empty.
- Fixed: Subset Particles by Statistic job no longer fails when manual thresholds are specified in the job builder.
- Fixed: Select Volume job now produces passthrough outputs consistently based on input connections.
Install patch? (y/n): y
Downloading...
Downloading cryosparc_master_patch.tar.gz...
Downloading cryosparc_worker_patch.tar.gz...
Patching... checking versions
Done.
Gathering worker info...
Patch worker xxx? (y/n): n
Skipped worker patch.
To patch workers:
1. Upload the worker patch cryosparc_master/cryosparc_worker_patch.tar.gz onto each worker node
2. Log into each worker node and run "cryosparcw patch"
e.g.,
cp -f cryosparc_worker_patch.tar.gz /opt/bioxray/programs/cryosparc2/cryosparc2_worker/cryosparc_worker_patch.tar.gz
/opt/programs/cryosparc2/cryosparc2_worker/bin/cryosparcw patch
Finishing...
Done.
Patch v4.7.1+250814 applied!
To complete installation, restart cryoSPARC with the following command:
cryosparcm restart
I see the “version” file in the master installation folder, but a “patch” file isn’t generated. It is there in the worker installation folder. I have restarted the cryosparcm service. I’ve also tried to launch a job, which fails with the error that the master and worker isn’t the same version.
Any ideas for how to fix this?
cryosparcm is /opt/cryosparc2/cryosparc2_master/bin/cryosparcm
total 100
drwxr-xr-x 3 anbog anbog 4096 Jun 19 16:12 .
drwxrwx--- 15 anbog anbog 4096 Aug 19 13:28 ..
-rwxr-xr-x 1 anbog anbog 76613 Jun 19 16:12 cryosparcm
-rwxr-xr-x 1 anbog anbog 761 Jun 19 16:12 gitmulti
-rw-r--r-- 1 anbog anbog 2065 Jun 19 16:12 minicurl.py
drwxrwxr-x 2 anbog anbog 4096 Jun 20 15:44 __pycache__
-rw-r--r-- 1 anbog anbog 2511 Jun 19 16:12 watch_restart_command.py
The installation has gone through patches and updates successfully many times before.
I’ve just tried downgrading to 4.7.0, then updating and finally patching again, but I still end up with only the worker patched and the master stuck at 4.7.1.
Thanks for posting these outputs. I am unfortunately still not able to tell what is causing the problems during patching.
Please can you run the following commands
csm="/opt/cryosparc2/cryosparc2_master/bin/cryosparcm"
cd $(mktemp -d)
(
$csm patch --download
ls -l $(dirname $(dirname $csm))/cryosparc_master_patch.tar.gz
bash -x $csm patch --install
$csm status
) 2>patch.err | tee patch.out
pwd # note path to retrieve out an err files
I manage an instance running the CUDA12 branch (v4.7.1-cuda12) and by mistake I patched it with patch 250814, but now I saw that I should have applied patch 250811 instead. What should I do now?
@daniel.s.d.larsson If you used the standard the standard CryoSPARC patch procedure, there is no harm in applying the (-cuda12-specific) 250814 patch to a v4.7.1-cuda12 installation.