Snapshot replication also makes use of the distribution agent to apply the
snapshot to the subscriber database. The way the snapshot replication works
is just like how snapshot is applied for transactional
When the snapshot agent runs, in both snapshot and transactional
replication, it posts sync commands into the MSrepl_commands table which the
distribution agent picks up and applies on the subscriber
These are the parameters you can use to filter the results of
sp_browsereplcmds [ [ @xact_seqno_start = ] 'xact_seqno_start' ]
[ , [ @xact_seqno_end = ] 'xact_seqno_end' ]
[ , [ @originator_id = ] 'originator_id' ]
[ , [ @publisher_database_id = ] 'publisher_database_id' ]
[ , [ @article_id = ] 'article_id' ]
[ , [ @command_id = ] command_id ]
[ , [ @results_table = ] 'results_table' ]
[quoted text, click to view] "combfilter" wrote:
> In article <Oq2WvQxJGHA.3064@TK2MSFTNGP10.phx.gbl>,
> hilary.cotter@gmail.com says...
> > what shows up in sp_browsereplcmds. I suspect what you will see there are
> > the commands which are used for the sync - i.e. to build the tables and
> > other objects on the subscriber, and the bcp the data there. IIRC the
> > distribution clean up agent does not clean these up until you are past the
> > retention period.
> >
> >
> why would the bcp data be in the distribution db? wouldn't it just push
> that across the net to the subscriber straight up as a .bcp file? why
> would a snapshot need to put anything in the distribution db other then
> "hey you need to push this .bcp across".
>
> ok so what you are saying is that snapshot replication DOES actually
> post transactions to the distribution db? I thought it just generated
> the .bcp, idx and whatever that 3rd file is and pushed those across the
> net with just a few instructions on where to put them.?
>
> sp_browsereplcmds times out for me. We have a lot of subscribers on
> this box. Is there anyway I can use that sp and just look at a certain
> db id?
>
> thanks.
