We, The Supporters, Want Metadata Backup/Restore

[Endymion 6]

Was the library.db built with this metadata already in place?  As I can delete my library.db and allow it to rebuild and appears to use those same ids.  Every scenario is unique in its outcome.

 

 

But a OS or folder level backup is your best option.   As you can restore a earlier version when issues arise.

 

So, does that mean I can copy my "library.db" over and it should work... where I don't have to scan again? Of course, having all the metadata in place. Where is that file? I have to do all this moving around in SFTP.

 

EDIT... never mind about where the "library.db" file is.

 

And as for the Title of this topic.  I as a "Supporter" would never use or need this option.

 

Lmao... and that is because you know this thing inside and out or just about. I would think, you wouldn't be a MOD for nothing. What about us dumb people (myself) or the ones that don't know the ins and out. It really shouldn't be complicated to back-up and restore libraries.

Edited September 23, 2017 by Endymion

[gstuartj 40]

So, does that mean I can copy my "library.db" over and it should work... where I don't have to scan again? Of course, having all the metadata in place. Where is that file? I have to do all this moving around in SFTP.

 

Again, that file is in the directory I mentioned several times before. If you have a /var/lib/emby-server directory, copy the entire thing. It will copy all databases and metadata.

 

This thread feels a little like Groundhog Day.

[Endymion 6]

Again, that file is in the directory I mentioned several times before. If you have a /var/lib/emby-server directory, copy the entire thing. It will copy all databases and metadata.

 

This thread feels a little like Groundhog Day.

 

oops... sorry, I didn't see your post. No, it is slightly different. That's mainly because the way Asustor gives permissions for folders (from my understanding). I meant to come back and correct myself on that.

Edited September 23, 2017 by Endymion

[Happy2Play 9780]

So, does that mean I can copy my "library.db" over and it should work... where I don't have to scan again? Of course, having all the metadata in place. Where is that file? I have to do all this moving around in SFTP.

 

EDIT... never mind about where the "library.db" file is.

 

 

 

Depends on if that library.db corresponds with that metadata.

[boxtor 3]

Depends on if that library.db corresponds with that metadata.

I think that's a problem.  What happens if there's a breaking change to library.db or corruption?

[ebr 16177]

except @@Luke said this:

 

 

https://emby.media/community/index.php?/topic/50425-32292-genre-cleaner/?p=481713

 

So it would seem only .nfo is enough is this case not an embryo folder restore?

 

That statement is in relation to only that one plug-in and only this one time.  

[boxtor 3]

Hi.  We offer you the ability to store the metadata alongside the media.  This is a very good way to be sure your metadata is always where you need it and allows it to be backed up with normal backup solutions as well if you wish.

 

We have not concentrated on a full backup feature primarily because there are many pure backup solutions available that could accomplish this so we are concentrating on the media-oriented features of the system at this time.

Manually added cast member information exists in the database.  There is no import/export for that.

[Happy2Play 9780]

Well breaking changes @@Luke tries to avoid. 

 

In testing I have notice deleting the library.db and allowing it to rebuild, it used the same itemids on my test library.  But every scenario is a unknown. And all my testing is on a Windows machine.

[Happy2Play 9780]

Manually added cast member information exists in the database.  There is no import/export for that.

 

From a metadata not saved with media standpoint that is correct.  But if saved with media that is not correct as it is saved within the nfo file.  But it all can be done manually via the database People table just like the manually backup/restore of userdata.

 

https://github.com/MediaBrowser/Wiki/wiki/Backup#how-to-backup-manually

[boxtor 3]

How do you migrate the people over from the old database to the new database?  It sounds like you would need to do a SQL query from old to new such as used in migrating user data.

[Happy2Play 9780]

How do you migrate the people over from the old database to the new database?  It sounds like you would need to do a SQL query from old to new such as used in migrating user data.

 

All trial and error for me since I have never dealt with databases.  But I just figured the command line out Luke has in wiki.

REPLACE INTO NewDB.People SELECT * FROM People

Note "NewDB" is the name you use when attaching the new database.

 

Steps

open backed up database

attach the new database and use what ever name you like but ensure to use that name in command above

run the command above

then close the database

 

@@Luke can these steps be added to Restore Manually in WIKI.

[boxtor 3]

@@Happy2Play

 

What about the items in TypedBaseItems?  That contains the crucial information that's manually entered for each person.  The only display information People contains is the name, person type and role (character name).

Either all the information would need copied into a new table or only the items associated with an entry in People would need copied.  Since TypedBaseItems contains more information than strictly People data, it's likely the latter.

[Happy2Play 9780]

The only real way to do this is via server folder backups if you are having issue.  Trying to pull other information besides userdata could just corrupt the database when the server tries to link people to their media items.  People are not independent of their media.

[boxtor 3]

I thought people were shared between media.  There are the names associated with each video.  However, there is common data (ie, DOB, overview, etc...) that resides in TypedBaseItems.  If I have problems with that data, the only way to recreate it is to manually enter it again.  This is why I was requesting a way to export that data and re-import.  Everything else has a mechanism for backup and restoring through the NFO files, which I heavily use.

 

It seems that the only way to back up this data is to copy everything to a new install via copying the library.db.  If I have problems with the library.db, I need to start over completely.