hey dan, thanks, that's what i was expecting (but hoping wasn't the case :)
i'm using ruby on rails for the app, so hopefully it will only hold connections to the db open long enough for each request? can anyone confirm this is true? if i just have to wait for existing connections to die off that's ok, but if i have to cold stop the whole rails app and restart, then that could be a little more disruptive... On 9/27/07, Dan Kennedy <[EMAIL PROTECTED]> wrote: > > On Thu, 2007-09-27 at 14:26 -0700, Cyrus Durgin wrote: > > hi, > > > > i'm using sqlite3 in a small project that will run on multiple servers. > > only one of the instances will be read-write to the database, the others > are > > all read-only. i understand that the recommended process for > replication on > > the read-write instance looks something like: > > > > BEGIN EXCLUSIVE; > > <copy database file> > > COMMIT; > > > > is it roughly the same on the read-only (destination) side? in other > words > > is: > > > > BEGIN EXCLUSIVE; > > <copy new file into place> > > COMMIT; > > > > the correct procedure? > > I think you will need to make sure that all SQLite instances have > closed the read-only database before copying over it. Otherwise all > the existing connections will continue to read the old data. > > Dan. > > > > > ----------------------------------------------------------------------------- > To unsubscribe, send email to [EMAIL PROTECTED] > > ----------------------------------------------------------------------------- > > -- Cyrus. <[EMAIL PROTECTED]>
