December 19, 2001 at 11:07 pm
Hi All,
We are planning to use log shipping (SQL 2000) to maintain our stand-by server. I have a few doubts regarding this
1. To use BCP i have to use bulk-logged recovery. But, what happens to my bulk inserted data. Will this be transferred to my standby-server.
2. If I use DTS for the purpose of uploading data, can i make use of full recovery model.
3. If I use federated databases should I have to log ship the second database also.
4. How exactly does the log shipping work. Does it do the necessary changes to the data at the stand-by server or does it run the transaction at the 2 server.
I would be very thankful if someone can provide some insights into this.
Regards,
Mitra
December 20, 2001 at 2:53 am
Hi Mitra,
I use log shipping as part of a solution to have a standby server running SQL 7 while we start running SQL 2000 in the live environment. A few points I've noticed:
Log shipping takes a while to initialise - one maintenance plan for each database - and tune, and starts with a most recent backup of the source database. After that the logs are shipped to the standby server and the transactions applied at the standby server based on a copy and load schedule. If you change the schema of a database or perform non-logged operations on the source server, you'll need to reinitialise the standby server. Unfortunately, with the way the log shipping is administered, this will probably mean deleting the maintenance plan and starting again. If you try to avoid reinitialisation and alter the schema or do the non-logged operation on the standby server before the next log is applied this won't necessarily work, but in some circumstances may, eg if you are just doing selects on the bulk inserted data.
So :
(1) my understanding on Bulk-Logged recovery is that it disallows point in time recovery but apart from that the data insert will be logged so it will be transferred to the standby server.
(2) dts bulk insert is much like bcp so I guess the first answer applies, even with bulk-logged.
(3) Am not too sure what a Federated Database is : definition I found is "databases that contain independently operated databases that own their own data".
This could mean a few things in the case of SQL - could you explain the organisation of your system.
(4) Transaction logs are backed up at the source server, copied to the standby server and then applied there.
Finally, once it's set up, there are alerts you can set up to see that syncronisation is taking place, and the GUI provides several different views of the history to manually check up on it. I set it up last week (1 day for 8 databases) and we hanen't had any problems yet.
Hope this helps.
Paul Ibison
Paul Ibison
Paul.Ibison@replicationanswers.com
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply