How to change database directory in cryosparc v2?

Dear all,
I have a small question, I’ve installed cryosparc correctly and worked well. But I put the database directory of cryosparc into a rather small hard disk, which bothers me right now. Does anyone know a way to change the database directory?
Thank you very much!

I have exactly this question too - my cryosparc2_database dir is taking up 88G and my home SSD is rapidly running out of space.

@Linhua, did you ever get an answer to this? Would it be safe (with cryosparc stopped) to move the database directory and then symlink it back to the original location, or is there a less kludgy way to do this?


I just re-installed my CS2 on ubuntu 16.04 this weekend (tempted by the 18.04 upgrade message showing at start up – turned out to be a trap. 18.04==CUDA9+GCC6, not CS2-friendly yet ).

I am still trying to figure out a way to get the old cryosparc2_database re-connected with the new installation.

1 Like

Thanks for the tip re Ubuntu 18! I have been tempted by the same message but have yet to succumb, now I will approach with caution


@apunjani any comment on this - is it possible to move the cryosparc2_database directory? I am nearly out of space on the disk where it is located, so if nothing else I might try moving it and then symlinking it back to the original location. Does this directory need to be on an SSD preferably - will there be much of a performance hit locating it on a spinning disk?


Hey @olibclarke, @Linhua, @ZhijieLi,

You should be able to change the location of the database by copying/moving it.

Please ensure your cryoSPARC database is backed up first: cryosparcm backup

Ensure cryoSPARC is turned off.
Copy/Move the directory of the database to a new location (note: it must be on the same node that the cryoSPARC master is on)
Edit cryosparc2_master/ so that CRYOSPARC_DB_PATH contains the correct path.
Start cryoSPARC

Also, the database doesn’t have to be stored on an SSD, as mongodb automatically caches data it needs in memory. Also, the database is used as a metadata store, so the storage medium will not affect the speed of jobs.

Hope this helps!

Fantastic, thanks @stephan, I’ll give that a whirl!


Is it possible to migrate the cryosparc v1 database in a similar way?

Also, in future versions would it be possible to write the database in each project’s folder (like relion does) ?. This would eliminate the pain of losing the database when the system drive crashes.


Hi @istv01,

The database used in older versions of cryoSPARC is not compatible with cryoSPARC v2.0+. Regarding data loss, please take a look at backing up and restoring cryoSPARC v2’s database here.

Thank you very much for your help!