When you build a new Microsoft Cluster Server cluster always be sure to fail over all the drives between all the nodes before installing SQL Server on it and deploying the cluster to production. The first time that you fail over a cluster from one node to another the disks can take a very long time to fail over.
If you don’t test this fail over ahead of time your first production fail over could be quite a bit longer than you expect. Granted I’m only talked 30-60 seconds longer than expected, but your are clustering the SQL Server for the most possible uptime, and 30-60 extra seconds could be a big problem.
Do in other words test, then deploy.
No tech blog post today.
I just want to wish everyone a happy new year.
This question has come up a couple of times on various forums on the Inter-webs so I figured I’d write up something a little more in-depth than I’d be able to do on a forum.
There’s no way to easily flip a standalone SQL Server instance to a clustered instance. However moving a standalone instance into a clustered instance isn’t that tough, but it will require some down time to the clients. Continued »
Windows 2008 introduced us to code mode. It was great, an ultra thin server OS, with little overhead as there’s no UI. It was as awesome improvement to Windows, unless you actually wanted to use any other Microsoft technologies. There’s no .NET support so you couldn’t run .NET services, or ASP.NET websites. You couldn’t install SQL Server, or Exchange. If you wanted classic ASP, or PHP then you were set. Pretty must the only real use for these was as as domain controllers and other infrastructure users.
Windows 2008 R2 core mode is actually useful. Continued »
When I was at the Windows 2008 R2 release there was a ton of new information that was talked about. There are all new management tools for Windows 2008 R2. In fact there is an all new methology within Microsoft when it comes to building the management tools.
All to often when people deal with parallel queries they see that some threads of the query have a wait type of CXPacket so they assume that the parallel execution plan is causing a problem, but it probably isn’t.
Everyone likes nice defragmented SQL Servers. However if you use Diskeeper to do so be very careful before you upgrade to the new version 10, or if you are going to upgrade to the new version 10 you’ll need to make a small change to your configuration on each SQL Server.
All to often I see people online saying that they detach a database then delete the log file to shrink it. This is a very dangerous operation which can leave you with a corrupt database that you can’t do anything with. Continued »
SAN replication is a very power tool which will let you get real time (or near real time) backups of your database on another location on your storage array, to another storage array at the local site, or to a remote storage array at another site.
However like all technologies there are some things which it is really good at, and some things which it can’t/shouldn’t be used for. (Everything here is going to be SAN vendor dependent. This includes features, modules you have to buy, terminology, and probably some other stuff I’m not thinking of.) Continued »
Proper network design is very important to ensuring the highest availability between your storage array and your server. This design is important weither you are using a Fibre Channel SAN or an iSCSI SAN. The basic design is the same. Continued »