Thanks, Reinout. I'll try to answer your questions and explain it a little
bit better.
I have created the two tables on each database and set their seed and
increment values then. The table on the publisher already has data in it,
so when I turn on replication I set the Article defaults to Keep the
existing table unchanged. This allows the subscriber table to stay at (2,2)
while the publisher is set to (1,2). I really don't make any changes to the
subscriber after initializing merge. This does work sometimes, but there is
nothing consistant that I have done to get it to work. It is more like
dropping and re-initializing the merge until the ID's decide to work.
I'm not too sure where the 9 digit numbers came from either. The tables I
was trying it on at the time already had several thousand rows in them so
maybe it had something to do with the fact that there were already ID's in
the table for the existing rows. I tried it on another table, a newly
created one a minute ago and it was giving single digit numbers on Server 02
and Server 01 started at 202. I might be able to live with that on new
tables, but it would be difficult to change the existing tables, there are
some tables where the ownership by ID, like what Team the information is
for, is set manually when data is entered. Basically, changing existing
table ID's from the current format, atleast on those where it is working,
would take quite a while to update all of the tables and rows using those ID
fields. I may end up having to rebuild it all, but I would prefer not to if
I can keep from it, my deadline is Friday. I have read a lot of BOL
articles and searched the knowledge base and msdn, but I don't know that I
have seen the article that you are talking about so I'll take a look.
Thanks,
Matt
> Hi Matt,
> Firstly:
> I tried to replicate (excuse the pun) what you are doing by changing the
> identity seed on the subscriber but I can't seem to see how you are doing
> it. What steps do you use to change the identity seed on the subscriber
once
> you've done the initial merge? The table schema is locked at the
subscriber
> once you've applied the snapshot to the subscriber.
> Secondly:
> I'm not too sure where your 9 digit numbers come from because the default
> subscriber identity range is 100 meaning that your publisher will have the
> range from 1-100 and your subcriber from 101-200. Are you using an INT as
> your identity column? Check BOL (replsql.chm::/repldata_078z.htm) for more
> information regarding changing of the default range settings.
> -R
> > Hi Reinout,
> > I tried using the built in identity ranges, but they were inserting
large,
> 9
> > digit, numbers and my boss wants to keep the id fields "pretty" and as
> > simple as possible. If there is a way to set the numbers that are used
I
> > would be willing to try it again and see if it could solve my problem.
> I've
> > been working on this for a couple of weeks, when time permitted, and
have
> > gotten the even and odd numbered ID's working on some tables after a lot
> of
> > work and no clue why it started working. It seems that if you're using
> the
> > NOT FOR REPLICATION option this should work fairly well and be an easy
> > setup. I can't understand why odd ID's are being given to inserts on a
> > table that is set to (2,2) for the increment and seed.
> > .
> > I'm not too worried about adding more subscribers because if we up the
> > number of our systems it will be a global increase and I'll have to
> redesign
> > the database anyway to accommodate the new complexities of the business.
> > Thanks
> > Matt
> > > Hi Matt,
> > > Is there a reason you want to do it this way? SQL2000 merge
replication
> > has
> > > built in identity ranges that it can work with so you don't have to
> manage
> > > the identity ranges by yourself. You might consider using these as,
even
> > if
> > > you got your method to work, it would be hard to add a second or a
third
> > > subscriber.
> > > Once again, if there is a reason for you doing it this way, let us
> know...
> > > Regards,
> > > Reinout
> > > > I'm using Merge Replication on SQL 2000 and having trouble with the
ID
> > > > fields. Server 02 is the distributor and the table ID column is set
> to
> > > Seed
> > > > 1 and Increment 2 so that it will insert odd numbers and on Server
01
> > the
> > > > table ID column is set to Seed 2 and Increment 2 for even numbers.
> When
> > > > data is inserted into the table on Server 01 it is given an odd
> numbered
> > > ID
> > > > instead of giving an even ID number. I checked the table design
after
> > > > turning on Replication and both ID fields are still set correctly
and
> > set
> > > to
> > > > the NOT FOR REPLICATION option. Does anyone know how to solve this
> > > problem?
> > > > Thanks.
> > > > Matt
by