identity management...

now as i have gotten the bugs out replication seems to work great (2 days and
no more conflicts), but I have a general comment. Microsoft documentation
stinks.
Wheres the Microsoft article on setting up replication, INCLUDING
SCREENSHOTS? Why did the defaults leave me high and dry with respects to
identity fields being primary keys and not reseeding properly. Does it not
make sense to have identity management turned on by default in merge
replication where identity fields are involved? Had this been the case I
would not have even experienced a hiccough in the transition to a replicated
environment. As it was not the case, I called PSS, paid for a support call,
and the agent I spoke with (Terry) never clued me into this fact, he only
helped fix all the primary key troubles (using DBCC CHECKIDENT which is
manually setting the range and manually maintaining it, which I was also not
made aware of). Any support call is a two way street, but I feel MS has made
a great product, poor documentation and in my case caused poor customer
relations.
pUmkInhEd
There are a few articles on the web on setting up replication, but they are
mostly independant. In general, the microsoft documentation doesn't have
screenshots. There are some articles here:
http://www.mssqlcity.com/Articles/Replic/Replic.htm and there is now Hilar'y
book which has loads of graphgics:
http://www.amazon.com/exec/obidos/tg...lance&s=books.
As for the automatic range management being turned on by default, this might
be a good idea in some cases, but I would rather have a prompt. It might be
that you want to use manual range management instead, but either way, to
have a prompt would be enough to make you realize that some type of
intervention was necessary.
You don't have to use dbcc checkident on each subscriber - have a look at
http://www.replicationanswers.com/ManualIdentities.htm for other ways of
manual identity range management. Using dbcc checkident will be ok for
ranges, but more complex algorithms used to ensure no overlap can't be done
this way, because the seed and increment will both need altering.
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)
|||Thanks for your response Paul, I guess the underlying solution is to educate
myself. Manual range management does not fit my solution since it seems to
require some type of user intervention to keep things going. I appreciate
your input and also your suggested reading material.
Thanks, Brandon
"Paul Ibison" wrote:

> There are a few articles on the web on setting up replication, but they are
> mostly independant. In general, the microsoft documentation doesn't have
> screenshots. There are some articles here:
> http://www.mssqlcity.com/Articles/Replic/Replic.htm and there is now Hilar'y
> book which has loads of graphgics:
> http://www.amazon.com/exec/obidos/tg...lance&s=books.
> As for the automatic range management being turned on by default, this might
> be a good idea in some cases, but I would rather have a prompt. It might be
> that you want to use manual range management instead, but either way, to
> have a prompt would be enough to make you realize that some type of
> intervention was necessary.
> You don't have to use dbcc checkident on each subscriber - have a look at
> http://www.replicationanswers.com/ManualIdentities.htm for other ways of
> manual identity range management. Using dbcc checkident will be ok for
> ranges, but more complex algorithms used to ensure no overlap can't be done
> this way, because the seed and increment will both need altering.
> Rgds,
> Paul Ibison SQL Server MVP, www.replicationanswers.com
> (recommended sql server 2000 replication book:
> http://www.nwsu.com/0974973602p.html)
>
>
|||Manual range management is manual only in the initial
setup, but there again, automatic range management
is 'manua;l' by the same rationale - you have to
configure the form ;)
Admittedly it is simpler to do automatic, and I only
would use manual for nosync initializations.
Good luck,
Paul

>--Original Message--
>Thanks for your response Paul, I guess the underlying
solution is to educate
>myself. Manual range management does not fit my
solution since it seems to
>require some type of user intervention to keep things
going. I appreciate[vbcol=seagreen]
>your input and also your suggested reading material.
>Thanks, Brandon
>"Paul Ibison" wrote:
replication, but they are[vbcol=seagreen]
documentation doesn't have[vbcol=seagreen]
and there is now Hilar'y[vbcol=seagreen]
/0974973602/qid=1107634413/sr=1-1/ref=sr_1_1/103-5571393-
2419806?v=glance&s=books.[vbcol=seagreen]
by default, this might[vbcol=seagreen]
a prompt. It might be[vbcol=seagreen]
but either way, to[vbcol=seagreen]
some type of[vbcol=seagreen]
subscriber - have a look at[vbcol=seagreen]
for other ways of[vbcol=seagreen]
checkident will be ok for[vbcol=seagreen]
overlap can't be done[vbcol=seagreen]
need altering.[vbcol=seagreen]
www.replicationanswers.com
>.
>