Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
denis_bda
Starting Member
2 Posts |
 Posted - 2003-02-13 : 16:46:01
|
| I administer a SQL server for an ecommerce company that specifically requires no failures or loss of data.We have our server hosted by a highly respected american hosting firm. Though, 6 months ago, we had a raid card failure that took down our server for a whole day and severly affected our image to our customers.I decided that it was time to consider moving our servers so I put into a motion to have our servers hosted by our sister company that is certified as an international hosting facility. I had the servers set up so that every possible point of failure was covered.Unfortunately, there was a power outage, and the UPS backups for the servers somehow failed and the servers were once again down for a day...I've decided that its time I come up with a completely new solution considering the problems I've had. The first issue involved a raid failure, that by SQL server's shared disk clustering concept is not prevented. The second, where power on all servers was lost, a single location cluster was the point of failure.What I'm wondering, is are there any multiple point clustering solutions that would allow me to place servers in different locations that can effectively maintain a live mirror of data.I've considered moving to an Oracle solution, but given the sheer complexity of Oracle, my company does not have the resources to hire on an expert in oracle to manage the system at this time. Besides, the sheer trouble of switching from SQL server to Oracle is a major task in itself.I'd appreciate any suggestions |
|
|
tkizer
Almighty SQL Goddess
38200 Posts |
Posted - 2003-02-13 : 16:50:37
|
| [url]http://www.microsoft.com/technet/treeview/default.asp?url=/technet/prodtechnol/sql/deploy/confeat/sqlha/SQLHALP.asp[/url] |
 |
|
|
Merkin
Funky Drop Bear Fearing SQL Dude!
4970 Posts |
Posted - 2003-02-13 : 16:55:17
|
| HiYou could look into log shipping or replication. Depending on how you set it up, your second server may not be instantly up to date, but you could do it so you only lost a few minutes worth of data. With this however, failover would not be automatic. You would need to manually point at the new database, or write some code to do it. Unless... you had a web server at each of these locations, and ran a load balancing device that could redirect visitors to the other location if one node dies.They are just some thoughts off the top of my head. Hope it helps a little.Damian |
|
|
|
denis_bda
Starting Member
2 Posts |
Posted - 2003-02-13 : 17:15:29
|
| Thank you for your responses. I will look into your suggestions. |
|
|
|
|
|
|
|
|
Non-shared disk SQL server clustering