Loading data for Virtual library only

[luciaisacat]

Hi! I hope this is the right place to post this question: I manage a very large Calibre Library and the calibre-server is often crashing because of memory problems. If this is a common problem, I wonder if it would be possible to hack the Virtual library in order to create a metadata.db.subset to be used by calibre-server without having it loading in memory the whole library data. Does this make sense? Thank you for your frank opinions.

[theducks]

You can set up access to a VL. But since the V in VL is Virtual, the full DB is used with a filter specified

[luciaisacat]

Originally Posted by theducks

You can set up access to a VL. But since the V in VL is Virtual, the full DB is used with a filter specified

Thanks to theducks for its reply... this is my point exactly: how to launch a server with a subset database (in order to use less memory) without disrupting (or massive copying) the main library database. My question is IF according to your opinion it could be worth to develop an option, within the VL feature, that would instruct Calibre to write a new metadata.db.subset which could be then used for launching the calibre-server. Thanks again!

lucia

[theducks]

Originally Posted by luciaisacat

Thanks to theducks for its reply... this is my point exactly: how to launch a server with a subset database (in order to use less memory) without disrupting (or massive copying) the main library database. My question is IF according to your opinion it could be worth to develop an option, within the VL feature, that would instruct Calibre to write a new metadata.db.subset which could be then used for launching the calibre-server. Thanks again!

lucia

I'm sure Kovid would accept a patch.
But most modern (supported by Calibre) systems have a ton of RAM.
Also 'serving Calibre' is not real intensive, like conversion, so any old PC capable of running a recent OS should do (Linux having a gentle footprint, seem to get more done with less CPU)