Adding New Column & updating bulk data in merge replication

Hi,
We have about 50 databases (SQL Server 2000, SP3) which are merge
replicated. We merge replicate about 100 odd tables in each of these
database. We need to add couple of columns in one of our major transaction
table where most insert/updates are being done. This table presently on
average has 5 lacs records.
During testing, we noticed that it takes about 60-80 minutes to add a
column in this table. Considering the # of database we have where the change
need to implemented, we will not be able to plan the upgrade without
production downtime. For upgrade 50 database it will take about 50 hours.
What are the options available in Replication so this can be done quickly
w/o any production downtime.
Adding to this, in one of the column we have added to the transaction table
, we need to update a new value. On testing we found that for 5 lacs records
it takes anyway between 2-3 hours. This takes roughly another 75 hours for
us to do this update after adding the new column in the table. How can this
be speeded up ?
thanks,
Soura
What's a lacs?
Basically there is probably no good solution for this. I would look at doing
a sync. Then creating my publication and then backing it up and restoring it
to all the subscribers and doing a no sync subscription, or I would look at
regenerating a snapshot and distributing it after you have made the column
change.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"SouRa" <SouRa@.discussions.microsoft.com> wrote in message
news:A19172BF-E8AC-4BB7-9FED-87EEC5353D89@.microsoft.com...
> Hi,
> We have about 50 databases (SQL Server 2000, SP3) which are merge
> replicated. We merge replicate about 100 odd tables in each of these
> database. We need to add couple of columns in one of our major transaction
> table where most insert/updates are being done. This table presently on
> average has 5 lacs records.
> During testing, we noticed that it takes about 60-80 minutes to add a
> column in this table. Considering the # of database we have where the
> change
> need to implemented, we will not be able to plan the upgrade without
> production downtime. For upgrade 50 database it will take about 50 hours.
> What are the options available in Replication so this can be done quickly
> w/o any production downtime.
> Adding to this, in one of the column we have added to the transaction
> table
> , we need to update a new value. On testing we found that for 5 lacs
> records
> it takes anyway between 2-3 hours. This takes roughly another 75 hours for
> us to do this update after adding the new column in the table. How can
> this
> be speeded up ?
> thanks,
> Soura
|||5 lacs is 500 K or 500 thousand i.e 500,000
lacs is primarily an indian unit of measurment. 1 lac is 0.1 million
"Hilary Cotter" wrote:

> What's a lacs?
> Basically there is probably no good solution for this. I would look at doing
> a sync. Then creating my publication and then backing it up and restoring it
> to all the subscribers and doing a no sync subscription, or I would look at
> regenerating a snapshot and distributing it after you have made the column
> change.
> --
> Hilary Cotter
> Looking for a SQL Server replication book?
> http://www.nwsu.com/0974973602.html
> Looking for a FAQ on Indexing Services/SQL FTS
> http://www.indexserverfaq.com
> "SouRa" <SouRa@.discussions.microsoft.com> wrote in message
> news:A19172BF-E8AC-4BB7-9FED-87EEC5353D89@.microsoft.com...
>
>