Subsonic Forum

Please note that the initial import of data from last.fm will take a while, roughly 30 minutes per 10.000 tracks. You can follow the progress meanwhile, but MusicCabinet features won't work until it's finished.

igbar wrote:1) why use PostgreSql? I have read a lot about it .. and it has been all good. Buy my question is why not use mySql? It does appear to have the majority of the market .. oh .. and I already have it installed for other products/project. Just wondering.

igbar wrote:2) Installed all the bits and clicked off the process to import and then read this
Please note that the initial import of data from last.fm will take a while, roughly 30 minutes per 10.000 tracks. You can follow the progress meanwhile, but MusicCabinet features won't work until it's finished.
Is it really not available until this process completes? I have 160K songs over some 26K albums and 10K artists. If it can only process 10K tracks at a time at 30/10K it will take many hours to get this all done. I realize that dealing with an external org (last.fm) does put some restraints on what can be done .. but couldn't the system be used with just the last.fm bits missing until complete. At least 98% of all my tracks have fully defined metadata.

igbar wrote:Also, when it stops doing the update/meta read, how do you get it to continue with the ones it didn't do? I think i missed that somewhere.

Thanks
Igbar

hakko wrote:

igbar wrote:1) why use PostgreSql? I have read a lot about it .. and it has been all good. Buy my question is why not use mySql? It does appear to have the majority of the market .. oh .. and I already have it installed for other products/project. Just wondering.

At the time when I was about to choose database, MySql had just been purchased by Oracle and I heard that test code that they were adding wasn't open source anymore. The future of MySql seemed a bit uncertain. I like how PostgreSQL is based in a strong user community, how they have always put reliability first (early MySql seemed to put speed over reliability to get attraction), and how powerful it is. How do you make indexed like searches in MySql, for example?

igbar wrote:2) Installed all the bits and clicked off the process to import and then read this
Please note that the initial import of data from last.fm will take a while, roughly 30 minutes per 10.000 tracks. You can follow the progress meanwhile, but MusicCabinet features won't work until it's finished.
Is it really not available until this process completes? I have 160K songs over some 26K albums and 10K artists. If it can only process 10K tracks at a time at 30/10K it will take many hours to get this all done. I realize that dealing with an external org (last.fm) does put some restraints on what can be done .. but couldn't the system be used with just the last.fm bits missing until complete. At least 98% of all my tracks have fully defined metadata.

hakko wrote:You can start streaming music from your directory structure the moment it's installed. It'll take a couple of hours to get metadata for all your music (artist biographies, genre categorization, top tracks etc), because last.fm has Terms of Service stating that we can "only" do 5 calls per second to fetch info and we're fetching a lot. So meanwhile, the system will work in a limited mode, without the joy of having access to last.fm metadata. Don't forget that last.fm is a free service, you're not paying a dime for this. I think that's pretty fair.

hakko wrote:

igbar wrote:Also, when it stops doing the update/meta read, how do you get it to continue with the ones it didn't do? I think i missed that somewhere.

Thanks
Igbar

viewtopic.php?f=11&t=10457