Category Archives: Installation
Continuing yesterday’s post regarding the licensing issues I discovered using the MAP Toolkit, I was able to resolve the issue without much difficulty. The license was showing in the MAP Toolkit as developer edition. We have volume licensing so I am sure now that wrong installer was selected at the time of install because we do not use keys as they are populated by the installer.
To correct the problem, we ran the installer on the server effected and selected ‘Edition Upgrade’ and selected the media. When prompted for a product key, we had to fool the installer with some trickeration. On a separate VM, we installed SQL Server using our Enterprise media which generated a key code in the box at the same point in the installation. We then typed that key into our real installation and proceeded with the ‘Edition Update’ thus allowing the installer to realize that we were licensed for a different version. This successfully updated our edition to Enterprise as this was the license purchased for this server.
However, after stopping and restarting SQL Server to get the new version to register we noticed that the SQL Agent was failing with no errors noted in the SQL error log or the Agent log. We ran the ‘Repair’ from the installation center (shown in figure 1) to make sure that there was no corruption with the previous edition. That did not resolve the problem. When running the agent from the command line we got a different error that pointed us to the login being incorrect. Somehow running the Edition Update locked out the AD account which is used only by this service on this server so a user could not have been to blame (or should not be able to). Once unlocking that account, SQL Server Enterprise Edition was running smoothly! Enjoy!
- The <database_name>_Restored database generates a few SQL Monitor alerts for all of the newly restored temporary databases that are eventually dropped after the DBCC CHECKDB is run against them. On our system we got up to five notifications per database:
- Database file size increased (Custom metric)
- Rate of write transaction/sec increased (Custom metric)
- Custom metric collection error
- Page verification
- Database unavailable
- When using a plain jane AD account to install and run the backups, you need to go into each individual database and add that user to the ‘db_backupoperator’ database role. This is not necessarily Red Gate’s fault as SQL Server does not have this at the instance level (I still do not understand why, but maybe I am just an idiot), however the Red Gate tool maybe should have warned me about this at some point instead of cryptic error message on failure, in my opinion. The Red Gate security model states that the account needs to be part of the sysadmin fixed server role which goes against best practices (if you go further down it does give a workaround like I suggested but I didn’t make it that far the first time I read the page, nor the second).
- There was a few servers that I had difficulty installing the agent onto unless I used an admin account even though they had the ‘log on as a service’ and sysadmin role on the regular account as well as permissions on all of the folders. I could then switch the service back to that account after installation and the service starts fine. They might have been missing logon locally. I will try that on the next server to test it.
I will keep this thread running as it appears there are some other issues. Enjoy!
I would love to upgrade my servers to SQL Server 2012. We’ve met with Microsoft and we discussed the licensing ramifications and I got really excited about upgrading after getting the OK internally. That was a couple of months ago. The one piece holding us up? Vendors!
We do have some SQL databases developed in-house, but being predominantly an Oracle shop, most of our SQL Server databases are vendor provided. We have set forth the initiative to do all new development solely in SQL Server for cost savings. Additionally, we have begun moving some of the vendor provided databases to SQL Server. However, those are also at the mercy of vendors. My latest project only supports SQL Server 2008, not even R2. This is very frustrating from an enterprise DBA standpoint. I want to move forward! Enjoy!
We have a morning “production control” meeting daily where key people report on the status of their systems as well as announcing approved change management items. Most days this is an exercise in repetition but it is necessary in order for administrators to announce changes to their systems that might affect other connected systems. For me, bringing down one of my SQL Servers can have a far out reaching effect upon many other systems.
For example, we have one particular server that houses all of the databases for our internal IT systems, such as VMWare Virtual Center and Citrix and a few smaller ones such as Blackberry Enterprise Server. This is a beefy server to power our infrastructure, however it is a major hassle every time I need to do some maintenance on it. Yes we have maintenance windows. However, for many of these systems the particular admin has to intervene before I can take their particular database offline. The usual cry is that “hey you cannot take that down on that day because I have X job running” where I say “no problem, how about tomorrow?” Then someone else chimes in with “no good, I have this process that runs on that day.”
My solution? Setup a meeting and coordinate the maintenance, which will happen come hell or high water. Cooperation is key. The database server must be patched and maintained, but without cooperation many other systems will suffer.
After my first presentation at SQL Saturday Pensacola, I am ready to do another and need to submit for SQL Saturday Orlando later this month. All of that is not a problem, I just don’t know if I should do what I did in Pensacola and expand on it because the audience seemed to enjoy it or forge a new path. I have good arguments for both, maybe I will submit both and see which one gets picked, if they get picked, of course. I have three weeks to decide, but I should not wait until the last minute.
The presentation that I did in Pensacola was the Seven Deadly Sins of the SQL Server DBA. The other one I am thinking about doing is just a straight up best practices install guide. I am very passionate about best practices and there are so many people who I’ve spoken with said that they wish they had known some of the best practices at the time of their install instead of correcting them after the fact. It also may help that I was asked to give the Seven Deadly Sins at the next Tampa Bay user group meeting this month, so we will see how that goes, maybe it was a one-hit wonder and the decision will be made for me? What do you think?
Does anyone read the EULA (End-user license agreement) during software installations? I know I do not, who can? Lawyers.
We were just informed today that before installing any software, updates, or patches onto our desktops or servers we must have approval from our legal department. So on patch Tuesday, I may have to wait a few weeks to get approval if the patch has a EULA thereby putting my servers in jeopardy of the legal department?
We have also heard that they would notify us of any requested changes to the EULA. Now that is funny! Let’s review this, an organization with around 600 employees dictating to Microsoft the wording in THEIR EULA. Is today April 1st? No, well then this will be interesting to say the least. I wonder if me commenting on this might bring legal action upon me?
The IT department has just been run aground by red tape. Enjoy!
Yesterday I started a new project to downgrade our two new SQL Server 2008 R2 clusters down to SQL Server 2008 clusters. The uninstall went off without a hitch as we removed both nodes and then removed the support tools. I find it interesting that when we went to install the 2008 server, there was still tempDB data files which prevented the new install from moving forward until we deleted them. I am not sure if the other system databases were there as the installer did not complain about those. In hindsight, I probably should have removed all of the directories and files before installing as a general best practice but I did reboot the server prior to the new install and thought it would be fine.
Because I do not spin up new clusters everyday (that would be a great job), I took screen shots during the initial 2008 R2 install to serve as a guide because I knew that I had a total of three clusters to build by the end of the year. In this scenario, documentation is an amazing thing (well it is amazing in most areas but most DBAs become complacent about doing it myself included). Originally I built the first two clusters back in January and since they were the first clusters that I had ever built I wanted to document it as I am responsible for many systems and quite honestly I would not have remembered the settings chosen on each screen. Having worked with SQL Server for sometime, I could have configured a stand-alone server in my sleep, but I was not as confident with clusters. My confidence is building at this point.
Now, it appears that building 2008 SP3 servers is almost identical to 2008 R2 servers from slipstreaming through the configuration for best practices. In a future post, I will discuss some best practice troubleshooting I did for these reinstalls. Enjoy!
Today we embark on a somewhat sad journey. A couple of days ago, I told you a story about a vendor and the miscommunication of specifications. Today the chickens have come home to roost. We must replace the SQL Server 2008R2 clusters with 2008 clusters. These were perfectly good clusters, fine tuned and ready to burst out of the starting gate and win the triple crown, well you understand what I mean. I’ve grown attached to these clusters as they were the first ones that I have ever personally built from the purchase order to production.
Today, we will uninstall both of the SQL Server nodes and then begin the reinstallation of the previous version. We will run the installer on the passive node first. In the installer go to the maintenance section and choose “Remove node from a SQL Server failover cluster.” After this is complete then we will go to the active node and go through the same process. Tomorrow we will look at the reinstallation of SQL Server 2008 SP3. Enjoy!
Today Microsoft announced the RTM for SQL Server 2012 with general availability set for April 1st! The SQL Universe and twitter #SQLFamily is all ablaze with excitement.
I personally will wait a few hours to let the servers catch a breath with some restored bandwidth. What about you?
Personally I will be setting up a Windows 2008R2 virtual lab running SQL Server 2012 and migrating the VMDK to my desktop and laptop probably. What are your plans? Do you plan to download 2012? Are you excited about Denali going RTM?
Yesterday I told you about my failed attempt to install SSMS for SQL Server 2008 R2 as a Citrix application and the failure that ensued. My apologies for making you wait a day for the solution, but I was rushed with another production issue yesterday and decided to break this into two posts for dramatic effect.
I found the following solution on several websites including Microsoft:
You need to change the following registry key locations:
Change the values of SP and SPIndex to 1 and SPName to SP1. The original values in my case were 0 and RTM, respectively. This is a strange issue because Visual Studio 2008 was never installed on this machine.
I hope this helps! Enjoy!