I need to set up log shipping for about 10 databases. The primary server is accessed by our website and the secondary server is a BI server. What I am struggling to figure out at moment is the best way to handle a “failback” situation once a failover has occurred. How do I get the primary server back to being the main server for the website?
Does anyone have any suggestions or best practices to handle a failback?
We are using SQL server 2008.
I think the recommended solution is to not failback instead switch roles as described here. I think they even recommend that the failover server should be named something like Prod2 instead of JustincasefailoverServer
But you could switch two times and your set where you want to be. Rationalize it by that you want to make sure that the server who failed should prove that it is capable of being up and running for a day or two.
Related
I'm developing an app using asp and vb.net. Hitting a db that is sql2008 r2. There's an internal app which sends an email to a customer. the email contains a link which the customer clicks on, and then the page load of that page updates a database sitting on our dmz. I'm trying to write a service then which will query this database at various times, and then, based on that result, fire off an email to an internal group. Originally this was set up to fire the email from the box on the dmz, however our NA doesn't like having port 25 open like that, so now I have to rebuild the app internally to query that database, so that the inbound email can be generated on an internal box.
SO... my problem is making the connection in Visual Studio (2012). When you configure the sql data source to a box inside the network, all you need is the name of the server, and you'll get the drop down populated with the databases. At first VS wouldn't see the server at all. We turned on "named pipes" on the server, and then I entered the server name as ip,80 (this is the only port the NA will allow open) and now it will see it, however, before the dropdown gets populated, I get an error saying "A connection was successfully established with the server, but then an error... an existing connection was forcibly closed by the remote host." I know sql normally runs on port, what, 1443? something like that? but if I do that, it goes back to not being seen.
Is there a way to configure the sql data source to see this server? I've researched for a couple of days, but generally the topics have been working the other direction, or related to sporadic issues, which this isn't. Our NA isn't much of a programmer, so he doesn't know much about my end, only that he seems sure that using named pipes is the way I need to get in... however, beyond enabling them on the server, I don't know much about them, or if VS can even use them...
thanks in advance.
(I've been coming to this site for a long time now for answers; this is the first time I've ever had post a question)
Wow, a MS SQL Server in your DMZ???
Short answer is to tell your NA he doesn't need to open port 25 for you to SEND an email, unless there is some part of the story I am missing.
The better answer, get that server out of the DMZ and create a web service. They are easy and can be made very secure.
This may be a dumb question, and based on the fact that googling has failed me I'm betting the answer is "no", but I thought I'd ask in case someone else has figured it out.
We're finally putting our website on a server farm, which means we can't use InProc session management. We're using SQLServer mode instead, but we had a situation where our SQL Cluster crashed. During this time, none of our newer web apps were able to load because of an inability to connect to the session database.
So here's the question: Is it possible to automatically fall back to a different session management (StateServer for example) or dynamically change the connectionstring so that we can use a backup Sql Server?
For now, our plan is to use DNS and if the main SQL Cluster fails, simply switch the DNS to a backup, but that's a manual task, and takes some time. We were hoping to have some sort of automatic failover.
I am afraid that there is no way. Also switching the session state mode would also make your application crash because users won't be able to find the data that was stored in their sessions. So an advice I can give you is the following: use a dedicated SQL server for the sessions, don't use the same server as the one serving your application data. And if you can, progressively start to update your application so that it uses less and less sessions, store very small amounts of data until you completely get rid of it. Make it stateless. Then your application will become very scalable.
In my current project, we have to create a website (ASP.NET MVC) which is likely to have sufficient load to demand a server farm. I understand that if server farm is used, session states must be stored on somewhere else such as SQL server database or state server.
After some experimentation, we are inclined to use the state server mechanism but the fact that it will have single point of failure, makes me nervous. Is there any method by which we can avoid "single point of failure" when using state server?
There is something called session state partitioning that you could use, in order to avoid a single point of failure. If this still doesn't suit you, then you might consider trying the ASP.NET Velocity project, which it looks promising even though it is in CTP stage only.
If you want full scalability and redundancy, then you should probably use a SQL Server Cluster.
sharedcache (http://www.sharedcache.com or http://sharedcache.codeplex.com) has an implementation for sessions, it's not released so far but people are using it.
You could set up SQL Server replication to another machine or use a failover cluster.
This could potentially be expensive but would make your database component more robust.
Technically, your web server equipment room is a single point of failure, as well as your network, etc. I wouldn't necessarily be more nervous about session state than any of those.
I currently have an asp.net website hosted on two web servers that sit behind a Cisco load balancer. The two web servers reference a single MSSQL database server.
Since this database server is a single point of failure, I'm adding an additional MSSQL server for backup. I would like to setup my web.config files to write everything to both MSSQL servers, but only read from the "primary" database server unless it is unreachable for some reason, at which point the backup MSSQL server would be used.
Is this possible via a web.config file setting, or must this be done in code? Thanks in advance for any help.
New Information:
I just wanted to add further information on this topic after researching it for the past several days - Microsoft TechNet has a good article title "Implementing Application Failover with Database Mirroring" (http://www.microsoft.com/technet/prodtechnol/sql/bestpractice/implappfailover.mspx#EMD).
This specifically covers the database mirroring feature in Microsoft SQL Server 2005 and the new new "Failover Partner" connection string keyword that allows you to specify two server/db instances in a single connection string.
The article is well worth a read if your interested in implementing this type of feature.
What you want is called "failover", where if one database fails your queries are automatically redirected to the other. This is acheived at the database level, not the application. There are a lot of walkthroughs etc for setting up failover clusters: here's one for SQL 2000, and another for SQL 2005. Basically, once you set it up, the primary database communicates all activity to the secondary one. If the primary fails, the secondary is (almost) up to date and takes over.
The servers form a cluster, and look like a single unit - similar to the way your load-balanced web servers look to the outside world. The backup monitors the primary, and if the primary stops responding, the backup takes over receiving queries. If you're Googling, try also looking adding the keywords "database mirroring" and "quorum".
Its a bit more complex than that. Does your webpage write to the databases? Or just read?
If they write, then you'll have to worry about keeping the 2 databases synchronized, probably using mirroring or log shipping.
But what you are (in essense) talking about doing is setting up a SQL cluster.
I've written a blog entry that shows how to setup MSSQL Database mirroring as well as how to actually utilize it from a managed code perspective:
http://www.improve.dk/blog/2008/03/23/sql-server-mirroring-a-practical-approach
Nice answer from "Rick" but I just wanted to add my 2 cents of information. Normally for a setup with failover without a lot of expensive equipment, I would set it up like that:
You can have your 2 SQL Server box waiting for request and have a third box low-end system with SQL Server 2005 Express as a "health monitoring". What that saves you is 10K$ for the box and one SQL Server licence. SQL Server Express (as in Free) can do the health monitoring between the 2 databases servers without any issues.
That is my setup :)
I've got a database server that I am unable to connect to using the credentials I've been provided. However, on the staging version of the same server, there's a linked server that points to the production database. Both the staging server and the linked server have the same schema.
I've been reassured that I should expect to be able to connect to the live server before we go live. Unfortunately, I've reached a point in my development where I need more than the token sample records that are currently in the staging database. So, I was hoping to connect to the linked server.
Thus far in my development against this schema has been against the staging server itself, using Subsonic objects. That all works fine.
I can connect via SQL Server Management Studio to that linked server and execute my queries directly. I can also execute 'manual" queries in C# against the linked server by having my connection string hook up to the staging server and running my queries as
SELECT * FROM OpenQuery([LINKEDSERVER],'QUERY')
However, the Subsonic objects are what's enabling me to bring this project in on time and under budget, so I'm not looking to do straight queries in my code.
What I'm looking for is whether there's a way to state the connection string to the linked server. I've looked at lots of forum entries, etc. on the topic and most of the answers seem to completely gloss over the "linked server" portion of the question, focusing on basic connection string syntax.
I don't believe that you can access a linked server directly from an application without the OpenQuery syntax. Depending on the complexity of your schema, it might make sense to write a routine or sproc to populate your staging database with data from your live database.
You might also consider looking at Redgates SQL Data Generator or any other data gen tool. Redgates is pretty easy to use.
One other idea - can you get a backup of the live database that you can install in development to do your testing? If its just data for development and testing that you seek, you probably want to stay away from connecting to your production database at all.
Create testing stored procedures on server B that reference the data on server A via the linked server. e.g. if your regular sproc references a table on Server B say:
databaseA.dbo.tableName
then use the linked servername to reference the same database/table on server A:
linkedServerName.databaseA.dbo.tableName
If server A is identical in its database/table/column names than you will be able to do this by some quick find/replace work.
creating a linked server from .NET doesn't make any sense since a linked server is nothing but a connection from one sqlserver to another server (sql, file, excel, sybase etc etc), in essence it is just a connection string (you can impersonate and do some other stuff when creating a linked server).
One Way is to create two connection strings and access the approperiate database when required.
Second option is create connection for Database A only and create a link server For Databse B in Database.good article, i really like it. I am doing a bit on research about Asp.net connection and i found also macrotesting www.macrotesting.com to be very good source. Thanks for you article.....
Regards...
Meganathan .J