Radicale becomes completely unresponsive when uploading large calendar

[HeptaSean]

When importing a large .ics file (moving a whole calendar with lots of old entries from another service to self-hosted Radicale), it needs around 20 minutes to process and becomes completely unresponsive during that time. The Nginx reverse proxy before that returns a 504 to the user trying to upload and no other users can log in.

Is that normal? Can anything be done to circumvent that?

[pbiering]

PUT events somehow resource intensive, have you watched system and disk load in parallel. The locking mechanism, which is at the moment not that very granular can cause additional issues.

Is the behavior the same for both caching methods (mtime+size vs. SHA256)?