#musicbrainz-devel

/

      • ruaok
        no, thanks for the reminder.
      • ocharles
        ok, outta here, back in an hour
      • ruaok
        one sec.
      • slaves do have indexes
      • I think I know why the edit search is broken. :(
      • the raw tables got imported into the musicbrainz schema, not the musicbrainz_raw schema.
      • :-(
      • reosarevok
        warp: couldn't you live somewhere where trains don't cost 30 euros for 2 hours?
      • nikki blinks
      • ocharles
        ruaok: that's correct
      • ruaok: they aren't meant to be in musicbrainz_raw
      • ruaok
        reosarevok: we could go to london where you can pay 30 EUR for a 15 minute ride!
      • ocharles
        ruaok: also, indexes aren't the huge bottleneck in the import, FKs and constraints are more costly. but I figured what the hell, might as well make it zippy :)
      • reosarevok
        :p
      • ruaok
        ah, interesting. then my dbdefs settings were wrong and the search might work.
      • ocharles nods
      • reosarevok
        (ok, true, not 6€, a bit more, but still)
      • ocharles
        I originally planned to put it in mb_raw
      • but luks said why bother, and I agree
      • but the upgrade ran ok as master?
      • ruaok
        ocharles: deleting bad FKs now.
      • so, still waiting.
      • ocharles
        ah, ok
      • when did that start?
      • ruaok
        Fks started 40 minutes ago
      • ocharles
        hum
      • seems worth running this deletion ahead of time then
      • could you email me a prod to do that tomorrow?
      • ruaok
        yes, plz.
      • and DB_READ_ONLY mode is true to its name now?
      • ocharles
        no, good point
      • ruaok pushes it onto ocharles' stack
      • ruaok
        now doing PKs.
      • yes, we need to do this cleanup before the put things in read-only