CryoSPARC

CryoSPARC (Cryo-EM Single Particle Ab-Initio Reconstruction and Classification) is a state of the art software to process cryo-electron microscopy (cryo-EM) data. It is designed as a dispatcher for cryo-EM workloads on a group of servers and workstations. RCC provides support for CryoSPARC on the Beagle3 cluster.

Installation and Setup

In order to use CryoSPARC on Beagle3, you need to request for a license from CryoSPARC and send it to RCC via this form. RCC will then setup your account and provide you with your account details. Each user will also recieve a dedicated base port for their account, usually ranging from 39100-39300.

Getting Started

CryoSPARC's GUI can be accessed via Thinlinc. After landing in a Thinlinc session, open a terminal window and run:

cryosparcm start

You may check the status of CryoSPARC at any point to learn CryoSPARC's status by using:

cryosparcm status

Now, to access the GUI, open a browser within Thinlinc and type the host machine name (midway3-login3) in the address bar along with the port number assigned to you. The hostmachine name is provided in the config.sh file in the master folder of your installation directory.

Example

You may type something like http://midway3-login3.rcc.local:39190/ in the address bar of the Thinlinc browser.

Info

RCC's CryoSPARC host machine name is midway3-login3. You are required to use this hostname at all times.

Troubleshooting

Sock Error

If the error message looks like, unix:///tmp/cryosparc-supervisor–6410667835282660811.sock refused connection (already shut down?), follow the steps mentioned below.

Run cryosparcm stop
Delete /tmp/*.sock file that belongs to your user account. The exact name of the the *.sock file will be in the error message.

Kill any interfering zombie processes that are still running. You can find the process IDs using:

ps -ax | grep "supervisord"
ps -ax | grep "cryosparc2_command"
ps -ax | grep "mongod"

kill <PID>

Run cryosparcm start to start CryoSPARC again.

Database Failure:

Kill the mongo processes. You can find the process IDs using:

ps -ax | grep “mongod” kill <PID>

Delete the .lock file at <cryosparc-install-dir>/db
Start CryoSPARC again using:

cryosparcm start

Database Spawn Error

If the error message looks like,

Starting cryoSPARC System master process..
CryoSPARC is not already running.
database: ERROR (spawn error)

then follow the steps mentioned below

Make a complete copy of your db directory as a backup and keep it safe
```
cp -rav db db_backup
```
Stop CryoSPARC
```
cryosparcm stop
```
Remove the original folder
```
rm -rf db
```

Kill zombie processes

ps -ax | grep "supervisord" (kill only the process that is running from your cryosparc install)
ps -ax | grep "cryosparc_command" (kill all the matching processes related to your cryosparc instance)
ps -ax | grep "mongod" (kill only the process running your cryosparc database)

Start cryosparc which will recreate the database
```
cryosparcm start
```
Stop cryosparc
```
cryosparcm stop
```
Run
```
cp -rav db_backup db
```
cd into the db directory that has all your original database files and run the following bash commands
```
eval $(cryosparcm env)
```
```
mongod --dbpath ./ --repair
```
Start cryoSPARC again
```
cryosparcm start
```