MSCS is not hard to set up and there are many available walk throughs on the Internet.
But now install SQL Server 2005 in failover mode.
Step back a sec, what is failover mode you ask?
It is a smallish cluster of 2 nodes, one active and one standby. This allows a decent amount of redundancy for your SQL server. It matches up with the KISS principle.
So you want to install SQL Server 2005 (and I will now reference it going forward as SQL2K5), so you log into your nodes to get it installed, insert the DVD, and start the walk through.
Huh? What is this reference to ‘Task Scheduler’ and the inability to run the remote install?
And now the frustration starts…
So you bring up your favorite search engine (I bet it is google!) and start looking around. Pages and pages of possible links on the screen, none of which give you any inkling to what is happening, yet other people are having the same issue. Or even better, you get to sites that look like they have the answer, if you wish to pay them to read the answer. That’s just not fair.
Now it is time to try out the knowledge base at Microsoft – it has a lot of good information. You enter in the message into the search box and get back 10-20 links of possible items, none of which seem to have anything to do with your problem. In fact, I’d say that of the first 2 pages of possible answers, not a single one of them had anything to do with the issue at hand.
Back to me and my experience…
Why won’t this thing install? What is the issue? The task scheduler service is running! I even looked, the job was there and scheduled to run in the same minute!
Yah, this has been a fun time, so let’s just start the whole cluster from the start, reinstall everything and try again, only to be stymied….again.
Off to search.live.com – it is from Microsoft, perhaps it can help. Again, type in the error message and look at the links that are shown. Wait, 7th one down, an offsite link to another place called SQL Monster but the link given from the search engine isn’t relevant. But this site looks like it has promise.
Now search this site for the error. 3rd link down I see the error message from someone else. I click. I read, and the answer is there.
Step back again – remember I said ‘log into your nodes’? Notice the plural?
You must not be logged into the second node when you install SQL2K5.
That’s right. The whole problem, hours of time, many searches in multiple engines, and it all comes down to the simple fact that you can not be logged in at all on the other node. Not on the console, not via RDC, not even a disconnected RDC session.
With that nugget of information SQL server installs cleanly (and relatively quickly) and my failover cluster is now ready for testing and deployment.
When I told my friend this, he responded: Yah, isn’t that just like Microsoft?
Right now I agree.