Ash cloud azure dnsmasq Domain Exim Linux oracle SQL Server

Its altogether LDAP

This past week I had two signifiant events. Well, three really. First I one a Innovation Voucher which will hopefully allow some research into a new product offering to occur. I say “hopefully” because so far all the NWDA has done so far is to ensure that my application meets some basic requirements. Then it will be up to some suppliers to show interest and see what happens from there.

Then I upgraded one my Linux Ubuntu machines to the latest Ubuntu. This took quite some time! The machine is quite a slow machine – its a Celeron 667Mhz. But it plays a critical role in the network as it runs DHCP, DNS, Backup, NAS, Samba and WINS. Overall the upgrade went through clean, but Ubuntu needs to ask a few more questions upfront – especially relating to config files. I would walk away from the machine, the display would turn off and then when I came back I was aprehensive to hitting a key lest I answer a critical prompt incorrectly. Maybe there is a way to keep the display on all the time. Another issue I ran into is the GDM login. I had turned this off. So users logging in at the console would login to text and then issue ‘startx’ if they wanted XWindows. The upgrade ignored that setting and reset the GDM login. I removed GDM from the startup but that still didn’t help. So I just removed all the Xstuff. And then I have to tell GRUB2 to go to a text screen.

Once all that is done and you’ve upgrade GRUB fully to GRUB2 the boot time is minimal. It really is fast!

The next event was changing ly local domain. I had a local domain which ended in .local. local is a public domain which, quite often would mess up my VPN users. And in certain versions of Linux, some utilities like ping will not work properly with a .local domain. So I decided to change it – to .localdom. My what a process. I had to change LDAP and all its config files and such. Surprisingly this was the easy part. Then all the ldap.conf files in all the Linux machines and the samba setup on the PDC, DNS settings (of course), my Exim setting – more on this below, Backuppc setting for e-mail domain.

Most things worked, except for the PDC. I determined through the logs that this was because the bind password was not reset. Once that was done things worked. That exposed a configuration issue with the way the NetBIOS browsing was being done. Apparently it is an absolute must that the PDC be the master browser. So back into smb.conf to correct. But nothing I did fixed the problem. Eventually I gave up and rebooted the machine. Things worked! Sigh. All that hassle and it was just a reboot needed.

Then, however a couple of days later I got complaints about e-mails bouncing. Upon enquiry with the receivers it seems the local domain name was being left in outgoing e-mails. The problem? The ldap query inside the exim config file which does that was not updated for the new base DN.

This week I’ve been looking into SQL Azure. This is Microsoft’s cloud version of SQL Server. Its quite basic to start but with Microsoft the best is usually yet to come. I’m sure this will be a very popular service. One that I may use in the future. I’ll be doing some of my data mining testing on it. Oracle does offer Oracle in the cloud, but its through Amazon ECC.

Analysis Ash cloud Changegate Consulting Data Mining Haworth oracle Trends

Would Data Mining be able to penetrate the cloud?

As I write this I am back in the office instead of on holiday. I should be on a ship in Red Sea relaxing by the pool. The Ash Cloud encompassing nearly all of northern europe has made life misery for many trying to get to work and return home. I ended up spending the weekend here and it was great until the obnoxious man that owns the Changegate car(clamping) lot in Haworth, Lancashire. Stay away from that one next time I go there and so should you.

This week I’ve been reading up on Oracle’s Data Mining offerings. Data Mining is the extraction of hidden predictive information from large databases. It is an absolutely powerful technology which companies should employ to focus on the important trend information in their databases. This information will help companies focus on the items which will increase or decrease their business performance. For example a company could use this information to trim down customer leads to only the most likely ones.

Data Mining has emerged now that there are ultra powerful computers and ultra powerful software like Oracle to deal with all the data. Two main areas of use are:

  • Automated prediction of trends and behaviors.  The most likely uses are with marketing and supply. Predicting the best leads for sales and when to buy stock in specific items.
  • Automated discovery of previously unknown patterns. The discovery of buying trends in such things as food or real estate.

Typically data mining is fed off the data warehouse. I you don’t currently have one, don’t worry. My colleagues at Chameeya Software Services can build you one. A modeller can then build, with the client’s help, an initial model where the answers are known. Then appy this model to the unknown areas. For example you have information on your current customers and you already know the answers about how much long-distance calling they use. But you don’t know this about your prospects, except that you have information about them, sex, age, location etc. You apply the model to the known customers, if it produces the correct answers, you then apply it to the prospects.

Oracle has now even started allowing customers to do their data mining in the cloud using Amazon’s cloud services (see link).

The bulk of work is performed using Oracle’s Data Miner tool. I’ll discuss more on this in my next blog.