April 30, 2006 | Comments: (0)
I said I would keep you up to date on how SP1 is going, so here's what I've gathered this week.
I could waste a lot of time telling you about the problems, but instead I've gathered a couple other blogs that go into some detail.
Here they are... However, the one thing I will say is that it doesn't seem completely safe to install it yet.
http://sqljunkies.com/WebLog/sqlbi/archive/2006/04/27/20615.aspx
http://blogs.conchango.com/jamiethomson/archive/2006/04/26/3870.aspx
http://blogs.conchango.com/jamiethomson/archive/2006/04/26/3877.aspx
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 30, 2006 06:41 PM
April 27, 2006 | Comments: (0)
Embarcadero announced today that they're partnering with DBE Software. They want to use their Schema Examiner to add functionality to ERStudio and RapidSQL.
What this does is analyze database models from a design perspective and provides recommendations based off of key elements. As with benchmarking, data modeling is one of my diatribes, and it's really a lost skill. Nobody really understands the modeling process anymore, nor do they really understand the relational concepts. Schema Exmainer's 'Teach Me' functionality will examine the model for you and tell you what's potentially wrong with it and why. Even for experienced modelers this can be a huge benefit because it's very difficult to reverse-engineer a model with 300+ tables and getting started is quite often the hardest part. Also, if you don't know the data very well it can also be very difficult to analyze the model.
Schema Examiner bridges this gap, and while it doesn't always catch every problem with a schema, it will take you a long way in getting started. I wouldn't just blindly make all the changes suggested, but I would use them as a reference as I went through the model myself. I find that schema problems are usually consistent, so whenever you find a problem in a table, the same problem usually exists in a lot of the others as well and Schema Examiner will be good at determining those system-wide performance killers.
In my opinion, this is the only real advantage ERWin users have had over ERStudio users for a long time. Especially with the new release of ERStudio, it is very quickly leaving ERWin in the dust. ERStudio already leaves ERWin buried in the usability category. It's GUI is much nicer, and the way it implements common functionality is much more user friendly. This new functionality will finally close that functionality gap that Embarcadero so desperately needed.
The major functionality list is below:
• Comprehensive Diagnostics and Reporting. Schema Examiner analyzes the schema’s data structure and generates graphic documentation of the entire database structure. The diagnostic can be customized.
• Design Discrepancy Isolation. The innovative “Show Me” facility isolates specific design issues in complex data models, eliminating the task of sifting through thousands of lines of code.
• Improved Productivity and Accuracy. A comprehensive set of diagnostics validates database design, recommends improvements, and automatically generates scripts to quickly execute design changes.
• Maximize Design Efficiency. Schema Examiner allows users to check and validate alterations to a database design, as well as simulate design changes before implementing them.
• Increase database performance by identifying violations of best practices that adversely affect performance. Suggested changes are scripted for easy implementation.
• “Teach Me” Facility. In addition to providing specific diagnostic and recommendations, the solution explains reveals the impact of design choices or modifications in light of Relational Theory.
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 27, 2006 07:26 AM
April 25, 2006 | Comments: (0)
Scalability Experts Sells DBStress
Idera has just released a new benchmarking utility called SQLScaler. The product has a useful data explosion feature that accompanies its benchmarking features.
I heard on the grapevine that they were about to release, so I contacted Idera a couple weeks ago about including them in a benchmarking piece I'm doing, and they denied that they had any such product in the works. My guess is that they're not up to par with the other products on the market like Quest Benchmark Factory and Mercury LoadRunner so they may be trying to keep a low profile... but that's just a guess.
I have managed to get a hold of a trial though, and I loaded it up on one of my servers. I was surprised to find that I had seen this tool before. It took me a minute to remember where, but it's actually a tool called DBStress, and it was written by Scalability Experts. I saw a demo of this product at the Scalability Experts office here in Dallas a couple years ago. At that point they were talking about shipping the product themselves. I suppose they just wanted to stay out of the software business.
Anyway though, I'm going to give this product a look and tell you anything you need to know.
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 25, 2006 11:27 AM
April 24, 2006 | Comments: (0)
Quest Spotlight Goes Enterprise
I got a press release in my inbox today that Quest's enterprise version of Spotlight for SQL Server has just been released. Now, I know that quite a few of you are aware that I'm not fond of Spotlight as anything more than a nice screensaver, so I'm going to give you tell you what you need to know about the new version.
First of all, it's really getting annoying that companies put the word 'enterprise' on a product without having any enterprise features to back it up. One of the problems is that an enterprise is so hard to define that they can get away with tagging anything with that moniker and get away with it.
I wouldn't exactly say that Quest is trying to pull a fast one by calling Spotlight an enterprise-level product, but I will say it doesn't live up to its name just yet. I believe the name represents the result they hope to achieve in the next couple years more than what the product stands for at this moment. It's a framework they're trying to achieve, and this is just the first release.
To date, the biggest problems with Spotlight have been performance related. These performance problems stem from bad design decisions at every stage of the game. Some of them have been fixed in this release and some haven't. For starters, Spotlight allows as many workstations to connect to the server as the server will allow. That means that if you have a server with severe performance issues you can connect with Spotlight to diagnose the issue... so can your partner, and your boss, and the PM in charge of the project. And what do you think that many admins connected to an already bleeding server will do for performance? I'll tell you what it will do... it will keep anyone from being able to get anything done, even in Spotlight, and the server will just die.
The new version of Spotlight actually fixes this issue by setting up a mid-tier that connects to the server, and all of the clients connect to the mid-tier instead. This is a much better way to do this, and it's good to see Quest is finally getting the message. Another problem Spotlight has always had is you could only view one server at a time in the console. The new version allows you to represent each server as a colored spot on the console, and if you have an issue, the color changes and you can click on it to see the details of the problem.
Those are the problems fixed by the new version, but there are still some of the old design flaws around. It's difficult to say exactly whether they're actually separate issues, or whether they're products of each other. See, Spotlight installs a work database on each server, and that database works almost entirely off of stored procedures that are poorly written and take a very large part in dragging down a dying system. I really can't count the number of times I've tried to use Spotlight on a server with severe performance problems, and I couldn't even get Spotlight to connect to the server because of the extra load it puts on the system. These issues seem to appear mainly in disk queues, and as it turns out, that's mainly when you see these issues with Spotlight. When the server is more disk bound, Spotlight seems to just exacerbate the issues in the way it stores info in its work database.
I know what you're thinking... how do I know the stored procedures are so bad? I mean, it seems that a product like Spotlight would encrypt its procedures to keep people from stealing their code. Well, as it turns out, that's correct... Quest does encrypt the stored procedures in the Spotlight work database. The problem is, there are plenty of procedures out there for decrypting stored procedures and they're easy to come by, and free. Just go to yahoo and type in something like "decrypt stored procedures in SQL Server 2000" and see what you come up with. You may have to refine your search a little, but you should find something very quickly. Then start decrypting your Spotlight work database and you'll see that they didn't encrypt the procedures to keep anyone from stealing them, but more to keep from being embarrassed.
In all though, Quest is aware of the performance issues surrounding Spotlight, and they're finally working to fix them. This enterprise release, though it may not be an actual enterprise release just yet, gives me hope that they may be able to bring some dignity back to Spotlight after so many years of pathetic performance. I am encouraged though.
I'm going to close with one more thought that I think is really important. I've talked to various officials at Quest about the Spotlight problem a number of times, and they almost always answer my attacks by saying they must be doing something right, because it's one of their most popular tools.
Well, I have no doubt that that's true. And knowing that not only has this extremely poor product been so popular for so long, and that Quest has known about these issues for years now, really tells me something about how much Spotlight customers actually rely on it for production issues... it also tells me something about Quest's commitment to its customers. Then again, if nobody's using it, who's there to complain?
Posted by Sean McCown on April 24, 2006 08:43 AM
April 20, 2006 | Comments: (0)
SP1 for SQL Server 2005 Yields Mixed Results
As most of you probably already know, Microsoft released SP1 for SQL Server 2005 this month. The service pack represents a few bug fixes as well as the production version of database mirroring. What you may not know however is that it's yielding very mixed results in production environments.
A lot of users downloaded the CTP when it became available and began testing it in their labs. The release was pretty stable, but the production release has failed to live up to the standards set by the beta. There have been widespread reports of SSIS errors and query anomolies, and other issues. I too tested the CTP and while I found no issues, I am reluctant to put it into production until I see how all this hammers out. My advice to you is to wait. Give Microsoft a chance to fix some of these issues before you put your production environment at risk. I'll be watching the progress of the situation and I'll be sure to let you know when I think it's safe to install.
Below here though, you'll find a copy of an email sent to one of the SQL Server usergroups. This has circulated quite a bit and it shows the types of problems you can expect if you install in production.
First impression ... this is the worst sql service pack for years.
After installing and testing the CTP-SP1 build 2040 on 4 development and test servers for 4 weeks with no errors, we decided to test the official
SP1 build 2047.
http://www.microsoft.com/downloads/details.aspx?FamilyID=cb6c71ea-d649-4
7ff-9176-e7cac58fd4bc&DisplayLang=en
The details:
Server A (32bit Win2003 and 32bit SQL2005): Installation runs smoothly, no errors on first sight. But then discovered, that sql is still reporting to be build 2040 (CTP-SP1) and we believe it, because it's the only test server which is fully functional (Integration Services) at the moment.
But
installer said SP1 Installation successful ... ???
Server B (32bit Win2000 and 32bit SQL2005): Aborted with Client Components installation. Found out that 630MB free on drive C isn't enough. After clearing 800MB on C, installation runs with no errors. After rebooting can't connect to Integration Services (HRESULT: 0x8002801D, TYPE_E_LIBNOTREGISTERED Microsoft.SqlServer.DTSRuntimeWrap) :-(
Server C (64bit Win2003 and 64bit SQL2005): First installation aborted with a failure on Integration Services, second try after a reboot says everything ok, no errors. But after a second reboot Integration Services service can't start. (The SQL Server Integration Services service failed to start due to the following error: The service did not respond to the start or control request in a timely fashion.) :-(
Server D (32bit Win2003 and 32bit SQL2005): Same as Server B, aborted with Client Components installation. But even after the successful second install has now a mixture installation with SP1 (build 2047) and CTP-SP1 (build 2040). After rebooting can't connect to Integration Services
(HRESULT: 0x8002801D, TYPE_E_LIBNOTREGISTERED
Microsoft.SqlServer.DTSRuntimeWrap) :-(
Well, after this experience we will stay on CTP-SP1 (build 2040) until there's a usable SP1.
Does anybody else tried the SP1 already? What are your experiences?
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 20, 2006 03:51 PM
April 20, 2006 | Comments: (0)
For those of you were putting your eggs in the Quickshift basket, you'd better have a backup plan because the rumor I hear is that they just went under.
Quickshift was a company that developed a memory compression technology that was supposed to revolutionize the database world by increasing server performance by orders of magnitude. I figured they had some problems getting things off the ground, but when I met with the CEO earlier this year, it sounded like things were going to be ok for them. I guess the SQL world just isn't ready to be shifted that direction.
There is a place for memory compression though. IBM uses memory compression to offset the overhead from their new row-level data compression in Viper. I think there will continue to be a use for compression technology, but as we see here, it's going to have to be used sparingly.
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 20, 2006 03:23 PM
April 10, 2006 | Comments: (0)
Microsoft Gains Foothold in BI
Microsoft has just acquired ProClarity, a company that develops a very popular BI visualization tool. This is going to prove to be a very good move on their part because currently Microsoft’s BI initiative is really weak on the client-side. Right now the only BI tool they have is Excel, and while it’s excellent for advanced charting and other business functions, it has failed to gain any real respect in the growing BI market as a serious analytical tool.
ProClarity already works very well with Microsoft’s tools with their tool Analytics Server, and other technologies, so this is a natural pairing.
I’m personally hoping to see Microsoft merge the two technologies in the future and really give BI users a complete toolset for working with the advanced information being thrown at them. I like to think that we can look forward to further integration into the Microsoft toolset by extending the functionality of ProClarity directly into Reporting Services, Office, Integration Services, and many more. With information coming at us faster than ever before, and from every direction, BI needs are becoming more unpredictable. Microsoft can even leverage this acquisition to further its objective of bringing data mining to the common man. Data mining has yet to make a strong presence in even the most serious BI shops because it’s still mostly too new and too complicated for end users and even DBAs to understand.
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 10, 2006 07:45 PM
April 04, 2006 | Comments: (0)
Why are the Backup Wars Important?
I've been talking off and on for a while now about the database backup wars, but what I haven't done is explain why they're important to begin with.
Well, for starters, there are two big issues: space and usability.
Space:
Databases are getting bigger and bigger, and as cheap as disk is these days, it's still very expensive to house all the backups you need to keep online. So there's a huge need for compressed backups, not only for the disk savings, but also to ease the increased network usage created by the backup process.
Usability:
There are other utilities out there that perform the same basic functions as the backup tools I've been talking about. Tools like ArcServe and Backup Exec can compress and encrypt backups as well, but they have one major disadvantage... you have to manage the entire process from within their own interface. The tools I'm talking about allow DBAs to manage the entire process from the SQL Server interface using the same tools and methods they use with the native SQL Server processes. They also have repositories that you can use to report on different stats including failures, operation length, etc.
As well, these tools allow you to convert the compressed backup files back to a native file format. So, what you get is the advantage of compressing your backups while maintaining control over the process.
Some of these tools also have advanced features that tools like ArcServe and the like don't. For instance, Quest Litespeed has object-level restore which is something that nobody else has.
Right now, everyone has pretty much mastered the backup engines and they all perform very well. So the war now has moved from speed and stability to price and features.
So, you should really care about how these wars come out because it's going to matter. And before you buy any of these tools, especially from Quest or Idera, always insist on a discount. There's no reason to pay full price for any of them because they're too expensive as it is.
Read my book reviews at:
http://www.ITBookworm.com
Posted by Sean McCown on April 4, 2006 09:28 PM
April 03, 2006 | Comments: (0)
I know a lot of you think that the database backup wars are finished, and for a while there, so did I. But, Quest is making sure that it's alive and well.
The war has been going on for a couple years now, and I'll briefly summarize the whole thing for you so you'll know a little about it.
1. Imceda Litespeed comes on the scene with backup compression and encryption.
2. Idera releases SQLSafe, a product that competes directly with Litespeed, and wages a very aggressive campaign against Imceda.
3. Red-gate buys MiniSQLBackup from Yeoh Ray Mond and turns it into SQL Backup.
OK, commentary so far. Litespeed had the entire market for a long time, then Idera came in and priced their tool much lower, so that forced Imceda to carve their niche in the higher end market and drastically lower their prices to compete with Idera in the mid-range and lower end markets. Unfortunately, Red-gate's appearance swept the low end market away from Idera, who frankly had a stronger hold on it than Imceda to begin with, so there's no loss there for Imceda. The problem was that even though Idera was a lot cheaper than Imceda, Reg-gate was significantly cheaper than everyone. This left Imceda with the higher market practically to themselves, while they were able to compete easily with Idera for the mid-range market. Idera really got the raw end because they didn't have a clear market. Red-gate squeezed them out of the low-end market effectively, while Imceda continually bashed them in the higher-end, and even some in the mid-range.
4. Walter Scott sells Imceda to Quest Software.
5. Everyone continues to add features and get small bugs worked out.
OK, more commentary...
Imceda selling to Quest was a very important move because it gave them a lot of breathing room for the price wars. They now had the backing of a much larger company with virtually limitless resources, which means they could lower their price to anything they needed to win any given contract. This way they could gain a stronger hold in the market and not have to worry about profits. It also meant that they were making almost nothing on very large sales.
6. Red-Gate releases an enterprise version of SQL Backup that contains a central repository.
That last one is the most important to our discussion today. It's because of this that Quest has been forced to re-work their pricing strategy.
Quest now has a multi-tiered pricing structure for Litespeed and it should really compete nicely with Red-Gate's low prices. What they've done is give you the ability to pay based on the size of your database. I'm having trouble getting specifics out of them, but I know the price can get very low.
The big question is, does this really put Quest in front of its competitors? It really depends on how you look at it. If you've got a fairly static environment, or at least very slow growth, then you could save a lot of money licensing this way. If you've got a very dynamic environment, you could spend more time upgrading licenses than actually doing DB work. In that case however, you'd probably just buy the high-end license to begin with.
Does Quest offer anything outside of features though? Well, for mixed shops the future is promising, but right now Quest has sat on the ball so long there's really no use counting on it. See, they've been promising Litespeed for Oracle, DB2, and MySQL for over a year now, but we still haven't seen anything. If Quest would step up, then there would be a huge advantage for mixed shops to choose Litespeed because they would have a cross-platform solution that would tie their entire environment together.
For now though, we wait.
Visit my book review site:
http://www.ITBookworm.com
Posted by Sean McCown on April 3, 2006 09:27 PM
TOP STORIES
WiMax OK for commercial useAgile mgmnt for small teams
Why developers avoid Vista
CBS to buy CNET Networks
Icahn's letter to Roy Bostock
Yahoo opens up Search Monkey
AT&T limits iPhone purchases
Silverlight gets put on Linux
Intel to develop PC with Alibaba
Cybercriminals can rent a botnet
ADDITIONAL RESOURCES
- Application Security: Threats and How to Counter Them
- Why Linux Threats Mean Business
- Virtualization: A Step by Step Approach to Success
- Is your smaller organization ready for High Availability?
- Is system maintenance doing more harm than good?
- Virtual Test Lab Automation: Manage development infrastructure
