Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chriss Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Real time replication? Sql's or mine? 1

Status
Not open for further replies.

IonelBurtan

Programmer
May 25, 2001
601
RO
Hi everybody,

I have a "replication problem" like that:
In a new system I must implement a structure of two mirrored SQl server with one identical database. Reason for this: system backup in real-time

Restriction:
- the databases must be identical at all times - real time, not from time to time.
- if one server fails the activity must continue with the other one, of course with a warning in all subsequent applications.
- when one lost server is back again, I must be able to get it mirror the running one WITHOUT stopping the running server.

What would you recomend? What probelms may I have (like id collisions, autonumbered columns etc.)?

Thanks in advance,
Any help apreciated

s-)

Blessed is he who in the name of justice and goodwill, sheperds the weak through the valley of darkness...
 
AFAIK the closest to your list of requirements is failover clustering (MSCS), but it requires shared disks.
 
I agree, cluctering is the closest thing that you are going to find. Anything else will be near real time, and the failover will not be automatic, and will require downtime to get back to the origional server.

Denny

--Anything is possible. All it takes is a little research. (Me)
 
What about Transactionale/merged replacation? Is it suitable for what I am looking for?

tnx

s-)

Blessed is he who in the name of justice and goodwill, sheperds the weak through the valley of darkness...
 
You could use the merge replication for this, however it would not be real time. It would be near real time.

Failover would need to be manual. In theory if the primary server crashed, you would be able to restore it with the KEEP_REPLICATION flag, when the restore was finished the distributor should feed the updated records from the subscriber to the publisher automatically.

But for this to work, I beleive that you would have to be doing log backups nonstop so that you always have the most current backup.

In practice you would need to create some downtime to backup the subscriber, restore it to the publisher and get the replication set back up, all while the system is offline so that the backup you restore is the most current version.

Denny

--Anything is possible. All it takes is a little research. (Me)
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top