You may remember we talked about Oracle's Unbreakable Linux the other day.
Anybody try to get on Metalink yesterday? The Register is reporting on the poor response time yesterday. Maybe Linux will kill Oracle...
Here's an interesting take from Dave Dargo...
Tuesday, October 31, 2006
Sunday, October 29, 2006
Firefox 2.0
I downloaded Firefox 2.0 this weekend to see what it was all about. I thought a couple pages of my regular sites loaded slowly, but that could just be my internet connection. Once of the first features I noticed was it's ability to detect a fraudulent website.
I received an email from a suspected ebay spoof. Sometimes I just click on the links to see how close they are to the real website. To my surprise, this message came up from Firefox:
Thursday, October 26, 2006
Will Oracle Kill Linux?
That's right, not will Oracle Kill Red Hat, but will Oracle Kill Linux?
There seems to be some buzz that Oracle's Unbreakable Linux is positioning itself against Red Hat's Linux. Oracle will be offering support on a version of Linux that they have basically ripped off from Red Hat.
There's no doubt in my mind that Oracle won't kill Red Hat. Linux is used for more than just running Oracle software. I should switch my whole enterprise of umpteen hundred Red Hat computers so I can run 20 Oracle servers more efficiently? I don't think so. Can you see a SysAdmin calling in to Oracle for support and having to wait 6 days to talk to Sandeep in some far reaching corner of the globe? Can't see it myself. Besides, Oracle needs Red Hat to continue to develop the platform so they can rip it off again.
The big question in my mind is will Oracle kill Linux? To successfully deploy Unbreakable Linux, Oracle is going to have to snuggle up to the hardware vendors in order to get Unbreakable Linux pushed out on their hardware. I would imagine that is going to tick off some of the proprietary vendors.
Or better yet, will Linux kill Oracle? Will Oracle get so distracted from it's core business of infrastructure software that the products go downhill further?
Only time will answer these questions. Of course, I don't have a yacht and a billion dollars, so what do I know?
At this point, I'd just be happy with a filesystem that doesn't reboot my box every 5 days.
There seems to be some buzz that Oracle's Unbreakable Linux is positioning itself against Red Hat's Linux. Oracle will be offering support on a version of Linux that they have basically ripped off from Red Hat.
There's no doubt in my mind that Oracle won't kill Red Hat. Linux is used for more than just running Oracle software. I should switch my whole enterprise of umpteen hundred Red Hat computers so I can run 20 Oracle servers more efficiently? I don't think so. Can you see a SysAdmin calling in to Oracle for support and having to wait 6 days to talk to Sandeep in some far reaching corner of the globe? Can't see it myself. Besides, Oracle needs Red Hat to continue to develop the platform so they can rip it off again.
The big question in my mind is will Oracle kill Linux? To successfully deploy Unbreakable Linux, Oracle is going to have to snuggle up to the hardware vendors in order to get Unbreakable Linux pushed out on their hardware. I would imagine that is going to tick off some of the proprietary vendors.
Or better yet, will Linux kill Oracle? Will Oracle get so distracted from it's core business of infrastructure software that the products go downhill further?
Only time will answer these questions. Of course, I don't have a yacht and a billion dollars, so what do I know?
At this point, I'd just be happy with a filesystem that doesn't reboot my box every 5 days.
Tuesday, October 24, 2006
I can see you!
My apologies for not posting anything for awhile but I’ve had some personal issues that I needed to take care of . So….. enough of that I’m back now and hope to contribute on a regular basis.
Let me seeee… We left off with the wonderful world of Spatial and me ranting about ESRI. I want to continue with the Spatial stuff but I wanted to share with you guys a more intriguing (at least in my opinion) turn of events. Even though this is totally out of my job description, in my shop we have to do what it takes to get the job done even if that means doing *shutter* development work.
We recently had a complaint from one of our customers that in certain parts of a country that his equipment couldn’t send or receive a satellite signal due to terrain obstacles and he would like to be able to see on the webmap those areas that are “blacked out” so he can avoid them. Well, since we’re all about keeping the customer happy I was tasked to “Make it happen”.
The first step in the process is getting the elevation data of the desired location, once you have the elevation, you find out which satellite your equipment is using. Once you know the satellite, you find out what the Lat/Log and height (location and distance above the earth) of it is.
When we have all that information we start piecing the puzzle together. We have to create physical objects in the database like a shape file (our satellite) and a Raster image ( our map) and input all the data that we gathered up (elevation, lat/log, height) once all the values are given we use a tool to calculate the black out areas. The example above, I’ve used an observation tower as an example. The red areas are not visible from the observers view but the green areas are. Get it?
This website gives you in real time all satellite names, and locations using a nice little java thingy. http://science.nasa.gov/RealTime/JTrack/3D/JTrack3D.html
Thursday, October 19, 2006
Taking the plunge
Within the last six months, I've gotten my enterprise off Oracle 8i and on to 9iR2. We were really behind the 8 ball (no pun intended) since the support window for 8i was quickly running out. We forged ahead and were able to get everything to 9i. Along the way, we upgraded some development systems to 10gR1 and put out a non-critical system in 10gR2. However, something was missing. We hadn't taken advantage of a great deal of the 9i features that didn't come straight out of the box (Optimizer enhancements, PL/SQL fixes, bug fixes, etc).
I didn't want the upgrade to 10gR2 to go the same way. I don't want a hard-and-fast deadline for 10g and I want to be able to take advantage of some of the bells and whistles that come with 10g to ease my management burden. My educational jumping off point is to be certified in 10g.
I'm a firm believer in certification for enhancing the individual's self-worth. That doesn't mean every certified DBA is worth something to a company, nor does it mean somebody certified is worth more than someone who is not certified. It simply means that I view the certification process a valid educational opportunity for the individual to benchmark his knowlege against a standard.
Now, it's not the full-blown-from-the-top certification path that new DBAs are going on. I'm certified in 7, 8, and 8i, so I'll be looking to upgrade from 8i to 10g through 1Z0-045. I know, kind of a cop-out, but I think this will give me an opportunity to try out some of the new stuff before I need it. I bought the book last week, so now I have to start studying.
Don't be surprised if you see interesting (to me, anyway) 10g thingies on the blog in the next few months.
I didn't want the upgrade to 10gR2 to go the same way. I don't want a hard-and-fast deadline for 10g and I want to be able to take advantage of some of the bells and whistles that come with 10g to ease my management burden. My educational jumping off point is to be certified in 10g.
I'm a firm believer in certification for enhancing the individual's self-worth. That doesn't mean every certified DBA is worth something to a company, nor does it mean somebody certified is worth more than someone who is not certified. It simply means that I view the certification process a valid educational opportunity for the individual to benchmark his knowlege against a standard.
Now, it's not the full-blown-from-the-top certification path that new DBAs are going on. I'm certified in 7, 8, and 8i, so I'll be looking to upgrade from 8i to 10g through 1Z0-045. I know, kind of a cop-out, but I think this will give me an opportunity to try out some of the new stuff before I need it. I bought the book last week, so now I have to start studying.
Don't be surprised if you see interesting (to me, anyway) 10g thingies on the blog in the next few months.
Friday, October 13, 2006
Apex Rocks
Two days ago, I knew very little about Apex (HTMLDB). I knew how to set it up and install it on the database side, but as far as developing something with it, forget it. There are other people using HTMLDB at my company and they have been getting good results with it. Since I didn't know that much about it, I kind of let them have free reign over things.
I am still supporting some reports that I did as a favor for a user a couple years ago. I don't mind since it's a real simple process, but the report relies on me running a query and dumping the results to a CSV file so I can send it to the user.
It's that time a year where I have to go through this process again and this year I decided to hand over the power to the user. From what I heard from other developers, it was a simple tool, so I gave myslef an extra week. I knew I could bang out hte CSV file in about 2 hours if I had to, so I basically had the whole week to work on it.
I started playing with Apex in the morning and in about four hours I had a basic report. In another day, I added some calendar pickers to let them enter a date range and some other dodads to give the user flexibility in how they wanted to filter the report. The best part was the "Spread Sheet" link that automatically downloads the report to a CSV file. Jeff - exit stage left.
I deployed it and gave the user the link in just under two days.
Apex Rocks!
I am still supporting some reports that I did as a favor for a user a couple years ago. I don't mind since it's a real simple process, but the report relies on me running a query and dumping the results to a CSV file so I can send it to the user.
It's that time a year where I have to go through this process again and this year I decided to hand over the power to the user. From what I heard from other developers, it was a simple tool, so I gave myslef an extra week. I knew I could bang out hte CSV file in about 2 hours if I had to, so I basically had the whole week to work on it.
I started playing with Apex in the morning and in about four hours I had a basic report. In another day, I added some calendar pickers to let them enter a date range and some other dodads to give the user flexibility in how they wanted to filter the report. The best part was the "Spread Sheet" link that automatically downloads the report to a CSV file. Jeff - exit stage left.
I deployed it and gave the user the link in just under two days.
Apex Rocks!
Wednesday, October 11, 2006
Passwords
One of the things about IT security that really irks me is passwords. As a user I need a password for system X and a different password for system Y. Not only that, but system X requires a password at least 6 characters long with at least one alpha character and one numeric character. System Y requires a password 8 characters long, with two numeric characters and I can never reuse the same password. To add insult to injury, System Y's password expires every 45 days and system X's password expires every 365 days. I just changed my password for system Y to something I know I'll never remember.
Now, I know what you are thinking: LDAP server. Centralize the authentication and authorization and you only need to supply a password once. That's all fine and dandy when I have control over the security, but not when system X is where I do my online banking and system Y is my brokerage account.
Things are changing in the financial world, and not for the better, IMHO. At some sites, I have to answer a personal question every time I login. Others, I have to choose a picture before I even get to enter my password. Others still, I need an RSA key along with another password. I think there should be a standard of authentication practices that your personal trading partners should have to adhere to. I've got so many passwords in my head, I can barely remember how to login to work. In the time I wrote this post, I've forgot system Y's password.
Now, I know what you are thinking: LDAP server. Centralize the authentication and authorization and you only need to supply a password once. That's all fine and dandy when I have control over the security, but not when system X is where I do my online banking and system Y is my brokerage account.
Things are changing in the financial world, and not for the better, IMHO. At some sites, I have to answer a personal question every time I login. Others, I have to choose a picture before I even get to enter my password. Others still, I need an RSA key along with another password. I think there should be a standard of authentication practices that your personal trading partners should have to adhere to. I've got so many passwords in my head, I can barely remember how to login to work. In the time I wrote this post, I've forgot system Y's password.
Wednesday, October 04, 2006
Truth and the DBA
I don’t tolerate lying at all.
I don’t even want you to spin the truth. Give me the whole truth and nothing but the truth and we’ll be fine. An article about conducting business in an ethical manner by Bud Bilanich at Trump University got my attention. It’s worth a good read.
When I first moved up the ranks from a team member to a team leader, I had somebody that worked for me that skated on the edge of the truth quite often.
“How’s project the upgrade project going?” I asked.
“Fine. I’m right on schedule”, she answered.
“Last time I did an upgrade, the JServ configuration gave me problems. How did that go with this upgrade?” I countered.
“No issues.” She replied.
OK, I guess they fixed that.
Two weeks before the big upgrade, I asked again if we were on schedule and she replied “Oh yeah, probably be done in a week.” So I sent a note to the users about the upgrade and how we’ll need people here to test on Sunday to make sure everything is fine. The users got their army ready for Sunday, upper management was notified since they had been breathing down our neck for getting this project done as well.
Monday before the conversion came and I ask for the new URL so I can look at the new software.
“Not quite done yet, definitely this afternoon.”
Hmm, something is sounding fishy here. I looked at the machine and the database wasn’t even up yet. I poked around some logs and saw that certain pieces were failing to come up for various reasons. Did a quick search on Metalink and saw a couple resolutions to the issues so I didn’t think they were too serious.
Tuesday and Wednesday I was out for training, but left explicit instructions that if progress wasn’t being made I was to be notified.
When I got back Thursday, I went to get a quick status.
“JServ doesn’t work, the Concurrent Managers keep dying, and Apache dies when you hit the login URL” was the reply.
Needless to say, the upgrade was cancelled. Upper management was steamed and since I was the project leader, it was my fault. That person no longer works for me.
Granted, it was my fault for not asking the right questions. However, if they had been truthful about their progress and struggles they would have garnered much more respect and I could portray an accurate picture of the progress to upper management. Their spin on the truth (or outright lies) caused my group to lose a lot of respect from the powers that be.
That’s one of the many reasons I manage the way I do today. As Tom Kyte puts it, “Trust, but verify.”
I don’t even want you to spin the truth. Give me the whole truth and nothing but the truth and we’ll be fine. An article about conducting business in an ethical manner by Bud Bilanich at Trump University got my attention. It’s worth a good read.
When I first moved up the ranks from a team member to a team leader, I had somebody that worked for me that skated on the edge of the truth quite often.
“How’s project the upgrade project going?” I asked.
“Fine. I’m right on schedule”, she answered.
“Last time I did an upgrade, the JServ configuration gave me problems. How did that go with this upgrade?” I countered.
“No issues.” She replied.
OK, I guess they fixed that.
Two weeks before the big upgrade, I asked again if we were on schedule and she replied “Oh yeah, probably be done in a week.” So I sent a note to the users about the upgrade and how we’ll need people here to test on Sunday to make sure everything is fine. The users got their army ready for Sunday, upper management was notified since they had been breathing down our neck for getting this project done as well.
Monday before the conversion came and I ask for the new URL so I can look at the new software.
“Not quite done yet, definitely this afternoon.”
Hmm, something is sounding fishy here. I looked at the machine and the database wasn’t even up yet. I poked around some logs and saw that certain pieces were failing to come up for various reasons. Did a quick search on Metalink and saw a couple resolutions to the issues so I didn’t think they were too serious.
Tuesday and Wednesday I was out for training, but left explicit instructions that if progress wasn’t being made I was to be notified.
When I got back Thursday, I went to get a quick status.
“JServ doesn’t work, the Concurrent Managers keep dying, and Apache dies when you hit the login URL” was the reply.
Needless to say, the upgrade was cancelled. Upper management was steamed and since I was the project leader, it was my fault. That person no longer works for me.
Granted, it was my fault for not asking the right questions. However, if they had been truthful about their progress and struggles they would have garnered much more respect and I could portray an accurate picture of the progress to upper management. Their spin on the truth (or outright lies) caused my group to lose a lot of respect from the powers that be.
That’s one of the many reasons I manage the way I do today. As Tom Kyte puts it, “Trust, but verify.”
Monday, October 02, 2006
The world of Spatial part I.....
I want to apologize even before I begin because there’s so much information that I feel I need to share with others that my topics may jump around.
Notice how I didn’t say “The world of Oracle Spatial”? That’s because there is an alternative in the land of GIS (Geographic, Information, System), you don’t have to use Oracle’s spatial module. The big dog in GIS is ESRI and if you use ArcSDE (component of ESRI) it has its own way of doing spatial stuff. Before I get down and dirty with the technical aspects of running and maintaining a spatial database, I feel that it’s important that you (the reader) know upfront you do have a choice of how you can mange your spatial storage.
*Steps up on soapbox*
The ESRI company started out as bunch of engineers who wanted to develop geographic software. The thought was great but, engineers have this idiosyncrasy about delegating work, they think they can do it all and it reflects grossly in their product. The first thing you will become distinctly aware of is that Oracle is kind of an afterthought in the eyes of ESRI, SQL Server is all that is holy with ESRI and the reason for that is because most of it’s customers run SQL Server. Why anyone would want to run an enterprise system with terabytes of data on SQL Server is beyond me (yes, I am Oracle biased). The second thing is Bind variables, they are unheard of, if you have a spatial database and you run ESRI get use to thousands of literal statements plaguing your shared pool . I have done battle for two years with these people and they just don’t get it. There are days where I really want to hop on an airplane, fly to CA and cause bodily harm to the development staff. End of rant.
*Steps off soapbox*
As far as using Oracle spatial vs ESRI “spatial” there are pros and cons of each, it’s up to you to decide which is best for your environment and skill set. I look at using Oracle spatial as kind of like “more moving parts”. The less “stuff” I have to deal with in a database the better, especially if it’s big. I figure if I let ESRI handle ESRI there’s less that can go wrong. I hang out over on ESRI’s forums and the number of Oracle Spatial problems are limited but when there are, they’re usually pretty bad and the question(s) go unanswered. As far as performance gained with Oracle spatial, the jury is still out on that one with me because I’m waiting to see what ESRI’s new release of 9.2 is going to be like. Apparently, they’re going to finally take advantage of the SDO_GEORASTER parameter. Right now they’re only using the SDO_GEOMETRY parameter which handles shape files. It’s like they developed a product, packaged it up, shipped it out, and forgot to put the CD’s in the box. Yeah, if you do use Oracle spatial (now) you get rid of your F and S tables but what’s the sense if you can only use half of the modules ability? It boils down to if you want to use Oracle Spatial when ArcSDE 9.2 comes out you’re going to have to completely drop your rasters and bring them back in (at least that’s the way I see it). I can see it now, managers across the nation giving birth to small farm animals when they find out they’re going to have to drop terabytes of data to take full advantage of Oracle Spatial because of ESRI laziness.
The last thing I want to do is go over anyone’s head when I’m on a roll talking about this stuff so if I mention something that you don’t quite get or want me to elaborate on, please speak up and I’ll be more than happy to pull the reins in sit for a spell.
Notice how I didn’t say “The world of Oracle Spatial”? That’s because there is an alternative in the land of GIS (Geographic, Information, System), you don’t have to use Oracle’s spatial module. The big dog in GIS is ESRI and if you use ArcSDE (component of ESRI) it has its own way of doing spatial stuff. Before I get down and dirty with the technical aspects of running and maintaining a spatial database, I feel that it’s important that you (the reader) know upfront you do have a choice of how you can mange your spatial storage.
*Steps up on soapbox*
The ESRI company started out as bunch of engineers who wanted to develop geographic software. The thought was great but, engineers have this idiosyncrasy about delegating work, they think they can do it all and it reflects grossly in their product. The first thing you will become distinctly aware of is that Oracle is kind of an afterthought in the eyes of ESRI, SQL Server is all that is holy with ESRI and the reason for that is because most of it’s customers run SQL Server. Why anyone would want to run an enterprise system with terabytes of data on SQL Server is beyond me (yes, I am Oracle biased). The second thing is Bind variables, they are unheard of, if you have a spatial database and you run ESRI get use to thousands of literal statements plaguing your shared pool . I have done battle for two years with these people and they just don’t get it. There are days where I really want to hop on an airplane, fly to CA and cause bodily harm to the development staff. End of rant.
*Steps off soapbox*
As far as using Oracle spatial vs ESRI “spatial” there are pros and cons of each, it’s up to you to decide which is best for your environment and skill set. I look at using Oracle spatial as kind of like “more moving parts”. The less “stuff” I have to deal with in a database the better, especially if it’s big. I figure if I let ESRI handle ESRI there’s less that can go wrong. I hang out over on ESRI’s forums and the number of Oracle Spatial problems are limited but when there are, they’re usually pretty bad and the question(s) go unanswered. As far as performance gained with Oracle spatial, the jury is still out on that one with me because I’m waiting to see what ESRI’s new release of 9.2 is going to be like. Apparently, they’re going to finally take advantage of the SDO_GEORASTER parameter. Right now they’re only using the SDO_GEOMETRY parameter which handles shape files. It’s like they developed a product, packaged it up, shipped it out, and forgot to put the CD’s in the box. Yeah, if you do use Oracle spatial (now) you get rid of your F and S tables but what’s the sense if you can only use half of the modules ability? It boils down to if you want to use Oracle Spatial when ArcSDE 9.2 comes out you’re going to have to completely drop your rasters and bring them back in (at least that’s the way I see it). I can see it now, managers across the nation giving birth to small farm animals when they find out they’re going to have to drop terabytes of data to take full advantage of Oracle Spatial because of ESRI laziness.
The last thing I want to do is go over anyone’s head when I’m on a roll talking about this stuff so if I mention something that you don’t quite get or want me to elaborate on, please speak up and I’ll be more than happy to pull the reins in sit for a spell.
Subscribe to:
Posts (Atom)
