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 |
|
MIK_2008
Master Smack Fu Yak Hacker
1054 Posts |
 Posted - 2011-01-17 : 06:31:49
|
| Hi there, Hope you’ll be having a good time. I have no experience what so ever on SQL server Clustering. But I have been given a task to implement SQL server Clustering for an application so that the application remains online even if the database server downs for some reasons. Currently, our production system has SQL server 2008 installed on Windows Server 2003 machine. There’s another machine on the same Network having SQL 2000 with Windows 2003, where I will need to install a separate instance of SQL 2008 side by side to the SQL server 2000. 1) Please let me know if Clustering is the correct solution for this task. 2) Can someone list each and every Step I will need to perform in order to achieve the highlighted task. This will help me to go in the right direction. (these steps will pin point the right materials to Google and to do the needful on asap basis)Also, can some one give ma any idea as how to achieve high Availability using SQL Server Express Editions! Many Thanks!~MIK |
|
|
dinakar
Master Smack Fu Yak Hacker
2507 Posts |
Posted - 2011-01-17 : 06:57:37
|
| "...application remains online even if the database server downs for some reason" is very generic requirement to suggest a high-availability solution. Do you currently have any HA/DR solution in place? What is the expected RTO/RPO?Link for Clustering: http://www.microsoft.com/windowsserver2003/technologies/clustering/resources.mspxDinakar Nethi************************Life is short. Enjoy it.************************http://weblogs.sqlteam.com/dinakar/ |
 |
|
|
Lumbago
Norsk Yak Master
3271 Posts |
Posted - 2011-01-17 : 07:29:55
|
| Clustering is quite expensive and requires very specific hardware. From your brief descriptions it does not seem to be the way to go. SQL Server Express has very limited high availability support..as far as I know the only option is Log Shipping which is more of a manual failover solution than HA.- LumbagoMy blog (yes, I have a blog now! just not that much content yet) -> www.thefirstsql.com |
|
|
|
GilaMonster
Master Smack Fu Yak Hacker
4507 Posts |
Posted - 2011-01-17 : 08:43:05
|
| For clustering you need minimum two identical servers (identical hardware, identical operating system, identical configuration), some form of shared storage, either a SAN or an iSCSI device and a minimum of SQL Server Standard Edition. It's not cheap.Mirroring's an alternative, but that also requires two servers and minimum of SQL Server Standard Edition.You can't directly implement log shipping on SQL Express, because of the lack of SQL Agent. It can be done, but it'll have to be custom scripts using windows scheduler. Not fun.--Gail ShawSQL Server MVP |
|
|
|
MIK_2008
Master Smack Fu Yak Hacker
1054 Posts |
Posted - 2011-01-18 : 00:38:17
|
| Hi thanks for your responses.@Dinakar, Currently we don't have any HA/DR solution in place and I am trying to implement one. But as far as RTO/RPO is concerned, we would ideally like to keep the recovery time as low as possible. However i don't know about all available HS/DR solutions supported by SQL server in their editions so I don't know how to give you the expected RTO/RPO info, as it seems that its highly dependent on the existing resources. I Would be happy if you could give me some details/links which helps me in defining an expected RTO/RPO, in light of our existing resources. @Lumbago and @GilaMonster: "From your brief descriptions it does not seem to be the way to go." Yes, I agree since I have not worked on this domain before so i am unable to give you more details except what we require and what resources we have in hand. And would expect you to suggest or provide me some links Which helps me in: first, Selecting a required HA solutions in light of its capability e.g. recovery/switching time etc etc.. And Secondly, implementation of the selected solution. As per your responses it seems that at this stage what I should care is of to select an appropriate HA solution. So please help me out in this. After selecting a solution i will ask for yours help in the implementation process :)Just FYI: Our application is using multiple databases located on the same SQL instance! Cheers! MIK |
|
|
|
GilaMonster
Master Smack Fu Yak Hacker
4507 Posts |
Posted - 2011-01-18 : 00:58:19
|
| "And would expect you to suggest or provide me some links"Tried google or Books Online?Entire books have been written on SQL HA/DR. There's lots and lots of material available. Start with Books Online. If you need this implementing soon, I would suggest that you get a consultant in to advice you. There's nothing worse than an incorrectly implemented DR solution.--Gail ShawSQL Server MVP |
|
|
|
Lumbago
Norsk Yak Master
3271 Posts |
Posted - 2011-01-18 : 03:27:56
|
| This link will lists the available features in each version:http://www.microsoft.com/sqlserver/2008/en/us/editions-compare.aspxYou say you have SQL 2008 and 2000 installed, are they standard versions? There are, as Gail says, numerous ways to ensure some kind of "high availablility" but it really depends on business requirements, money, and the amount of automation. Clustering and mirroring can provide automatic failovers, where mirroring is the cheapest of the two (and usually the best in my opinion). Other options, which are not really regarded high availability solutions, are log shipping and transactional replication. Both are very cheap but failovers are manual and there is some potential for data loss. HA for SQL Express is, as mentioned before, very limited but if you have sql server agent running on another server I think you can use that to do your backups for use with log shipping.- LumbagoMy blog (yes, I have a blog now! just not that much content yet) -> www.thefirstsql.com |
|
|
|
|
|
|
|
|
Clustering