The Informix User Group of Northern California are having their first meeting tomorrow Friday May 16 from 9:30 - 2:30.

Here is the agenda:

• Introduction (9.30 - 9.45) - Anup Nair, Technical Product Manager
• IDS Road Map (9.45 - 10.30) - Cindy Fung, World Wide IDS Product Manager
• IDS 11.50 Features (10.35 - 11.35 )- Keshava Murthy , SQL and Extensibility Architect
• Security and Compliance: IDS with Centrify Direct control (11.40 - 12.00) - Corey Williams, Director of Product Management and Phil Austin, Regional Sales Manager (Centrify)
• Lunch (12 - 12.45)
• MACH 11 PHASE 2 (12.45 - 1.30) - Ajay Gupta, BAR/HDR Architect
• Database Administration in IDS 11.50 (1.35 - 2.05) - Suma Vinod, Senior Software Engineer (Advanced Support)
• Demos (2.05 - 2.30)

The meeting is held at:

IBM San Jose
4400 North First Street, Suite 100, Room 1066
San Jose, CA 95134 (directions)

Most importantly lunch will be provided, so if you're in or around the Bay area take an extra long Friday lunch break and go and see what's new..

The folks over at Information Management consumability would like you to take the new Information Management Consumability Survey where you get a chance to say what you like and don't like about the Information Management products you use. This feedback gets a lot of attention and your comments really do make a difference.

I know what you're thinking, our products are already so wonderful how could they be improved? I know, but give these consumability guys a break, we don't want to put them out of a job.

On a similar note we had some very useful user feedback sessions at the IIUG conference, where in exchange for free t-shirts conference delegates would share their thoughts on the usability of our products. I attended the IDS Install sessions and we received some good suggestions which are being converted into feature requests.

This week I'm teaching an IDS Windows internals and debugging class in Lenexa, which is a follow-on from last week's class, and a team effort with tech support and development engineers helping to write and teach modules. As most of the material has been written for the first time it is fairly hectic, with the class schedule being decided very much at the last minute.

In the mean time, the IBM Informix Dynamic Server v11.50 Information Center is now up and running, in time for the imminent Cheetah 2 GA. The front page includes links to release notes and new features.

A few stragglers from the IIUG conference are camping out at the Doubletree for various meetings this week, and the consensus is that the first IIUG conference for many years was a great success as well as being fun. Good job Stuart and team. The sponsors I spoke to appreciated having such an Informix focussed attendance at the booths too.

If you were at the conference this week there is a good chance you'll see yourself in the IIUG Conference 2008 photos, as enigmatic Frenchman Jean-Georges Perrin was very efficient with his camera.

The IIUG conference is at an end, and for the rest of the week I'm busy helping with an IDS for Windows Administration and Support class taking place in Lenexa.

Long time Informix guru Jacques Roy has just announced a new developerWorks blog: Informix and Computing. In it he plans to start a dialog with the Informix community, discussing topics such as business problems and approaches to solving them, what Informix programmers are trying to accomplish, and if I know Jacques expect some great articles on extensibility too. He is also welcoming your comments and questions (which reminds I need to switch comments back on for this blog now the spam has died down), so take a look and let him know what's on your mind.

This week is IIUG conference week in Lenexa Kansas, and in this morning's keynote speech Dr. Arvind Krishna, VP for Data Servers, officially announced IDS 11.50, also known as Cheetah 2. Available in open beta program for a while, but until now closeted under NDA, we can finally talk about the new features in Cheetah 2. Here is a quick summary of what's new..

Administration

• Cluster management API Start/Stop, Add/Remove Configure
• Automated Update Statistics
• Configuration Wizard
• New Performance Graphs
• DRDA Config during Install
• Read-only access added to Open Admin Tool
• Row Versioning
• Wizard for creating new “tasks” and “sensors”

Application Development

• Dynamic SQL in SPL
• Additional IS [NOT] NULL support
• J/Foundation JRE 5.0 support
• DRDA enhanced Data Studio
• Tool to visualize explain output
• Basic Text Search support for XML and Xquery
• XML Transformations (XSLT) support
• BIGINT and BIGSERIAL

Security

• Single sign-on via Kerberos
• SSL Encryption

Mac OS/X Support

Continuous Availability

• Update any node in a cluster
• Connection Manager for balancing workload across nodes
• Policy-based Connections
• Arbitrator for flexible failover handling

The announcement is also covered in CNN Money: IBM Helps Clients Reduce Data Management Costs With New Informix Dynamic Server.

It is great to be at a conference that is all Informix all the time, with around 350 Informix users from around the world attending 4 different technical tracks, and as always good to catch up with people you know mostly through email and newsgroups. The organizers of this conference know how to cater for engineers, and load us up with sugary snacks every afternoon to help with those drowsy end of day presentations and dreary blog posts.

The Cheetah 2 open beta (IDS 11.50) started back in February, and since then clients and partners have been busy downloading it on various platforms and testing the new features, but not saying much about them, as all downloads were subject to a non-disclosure agreement (NDA). As a blogger you can bet I really enjoy being subject to an NDA..

As the IIUG conference and the Cheetah 2 release date approaches you'll start hearing much more about this new version of IDS and its enhancements. Next week on the final day of the conference (April 29) there is an opportunity to learn about the Cheetah 2 new features directly from the senior architects. In this eSeminar: Get the Scoop on IBM/Informix Cheetah 2 - from the Product Architects, the Informix Lab Director Jerry Keesee and 4 of the 5 Informix Senior Technical Staff Members (STSMs) Kevin Brown, Jonathan Leffler, Scott Lashley, Madison Pruet, will talk about Cheetah 2..

What's the latest in OLTP Data Server technology? Should you download the open beta version of Cheetah 2, IBM Informix Dynamic Server 11.5? What will Cheetah 2 offer and how will it impact your business' bottom line?

Join us for this live, interactive eSeminar roundtable discussion, directly from the IIUG Power conference in Overland Park, Kansas. Our discussion will be led by IBM Informix Lab Director Jerry Keesee, and will allow you the chance to interact with the IDS 'Power' team of architects.

They will give you a recap of the IIUG Power conference, but most importantly they will answer your questions - so bring them on and take your chance to provide feedback to the IDS experts.

You can register to attend the seminar here.

For anyone going to the conference, see you in Kansas.

It's always good to see some of the IDS core high availability and performance features like HDR, ER, fragmentation, PDQ used together in an innovative new solution where. A press release today from Jauka highlights IDS as the key to its new Content Fulfillment and Transaction Services for Mobile Media - Jaduka Launches Content Fulfillment and Transaction Services for Mobile Media.

At the heart of Jaduka's new content fulfillment and transaction services is its award-winning IBM Informix Database Solution (IDS), including high availability data replication, enterprise replication, fragmentation, parallel data query, and other proprietary functions that enhance stability and performance. Jaduka's IDS provides instant access to hundreds of millions of end-user accounts and the existing capacity to manage billions of transactions per month.

In a recent IT-Director.com article entitled Informix seeks developers, David Norfolk, a senior analyst at Bloor Research shares his thoughts on an "alternative" database, namely Informix Dynamic Server, and how it is looking increasingly attractive to application developers.

With quotes from the Informix lab director Jerry Keesee, there are some notes on IDS positioning as a database for "high-performance OLTP, especially in embedded (integrated) environments — because it is scalable, reliable, easy to manage and performs well".

The main focus of the article is that IDS "doesn't appear to be a dead-end as far as development goes." (Not a dead-end, I'll suggest that one to Marketing.) Evidence cited includes new developer friendly features such as the SQL Admin API, the free Developer Edition, the MAC OSX port, the Data Studio "Application Lifecycle Management" (ALM) environment, and the Solid acquisition with its expected synergy with IDS.

The article can in some ways be seen as a follow-up to an earlier IT-Director analysis by Philip Howard: IBM and Informix, where the IDS 11.10 release and some of its new features were highlighted, and the message conveyed that the future of IDS looks promising with IBM's increased focus on information on demand.

Here is a reminder for tomorrow's chat with the lab with Jerry Keese and John Miller III..

Tuesday, April 8th, 2008

8 AM PST, 10 AM Central, 11 AM Eastern, 4 PM London, 5 PM Paris

Optimizing IDS Performance

In this Chat, some of the more popular onstat options for performance issues will be covered, showing how they have evolved throughout the history of IDS. With this as a backdrop, guidance will be given to a performance-minded administrator on how to use the improved onstat information to get the most out of IDS. Our speakers will be Jerry Keesee, Informix Lab Director and John Miller iii, Senior Technical Staff Member.

You may RSVP for this Chat at: https://ww4.premconf.com/webrsvp/register?conf_id=9943887 . For RSVP assistance by phone, you may call 1-800-289-0579 and refer to the call with Confirmation Code 9943887.

Questions, comments: lspina _ at_ us.ibm.com

My erstwhile colleague Johann Backlund writes from Sweden:

...
I thought I would take a chance on guessing your email address to send 
you some of my experiences regarding IDS shared memory allocation and 
DLL's ...maybe something you could include in your blog series on the 
topic (part 1, part 2, part 3). 
I had planned to send this as a comment to your blog, but since comments 
are disabled...(hint, hint ;o)...

I'm working with a customer that just took the big step from 7.31 to
10.00. The migration went fairly well, but after a week or so they got
problems with onstat, onmode and oncheck. They got the standard error:

shmat: [22]: operating system error
MapViewOfFileEx: w32ec=487 at nt_shm.c:663
Unable to attach to shared memory.
Invalid argument

and this seemed to occur when the total shared memory size had grown to
around 1.7GB. They are running 10.00.TC6X2 with /3GB turned on.

We tried applying the KB913409 hotfix and ran this with both 10.00.TC6X2
and 10.00.TC7 (because of the fix for IC52767), but we could still
reproduce the problem.

After some experiments, I managed to pinpoint the exact segment causing
the problem and compared the set of DLL's loaded by oninit.exe with the
DLL's loaded by the other utilities. I found that there were two DLL's in
the address space occupied by that segment: WINMM.DLL and PSAPI.DLL.

By increasing SHMADD to 34000, that particular address space wasn't used
by IDS and this allowed the customer to allocate all of their memory
without causing problems for onstat et al.

So, this problem can actually occur even though DLL's are loaded in the
correct address space and the best way to avoid it is to set SHMADD to a
higher value to make it stay away from those small areas in the system DLL
address space. Also, I think Informix should consider to:

  * make oninit.exe load the same DLL's as the shared memory based utilities.
  * have oninit.exe stay away from the system DLL address space altogether
   (not sure that Win32 API allows this in a simple way).
  * add documentation on the problem with suggestions on increasing SHMADD
   - maybe even set the minimum value to a higher value than now (like 64M).

...

cheers

Johan

Thanks Johan.

About disabling comments. Unfortunately I had to disable comments on the blog the other weekend after getting 150 spam comments in the space of a couple of hours. Each one had to be manually removed. DeveloperWorks has since installed new a captcha mechanism, so I might re-enable them soon when I'm feeling brave. Comments have been temporarily switched off on IDS Experts too.

What Johan's comments show is that there is more than one approach to optimizing shared memory usage on the 32-bit Windows platform. You can rebase DLLs out of the way, but tuning the size of shared memory segments can be equally useful. Thanks to Microsoft not seeming to place a very high priority on DLL base address standards this problem needs to be revisited with almost every operating system service pack. At least 64-bit Windows gets over this problem with 8TB of addressable space.

The open source Ruby/Informix driver developed by Gerardo Santana has been updated to include INTERVAL data type support, along with Ruby 1.9 compatibility and other enhancements and bug fixes. Here is Gerardo's announcement in full:

Update 4/3/08

Unfortunately, one of the structural changes introduced in 0.7.0 was
not done right and caused that the name of Informix resources were not
unique.

This is fixed in 0.7.1, along with a documentation bug, and an
optimization that speeds up Database#columns (considered a bug also,
hence its inclusion in 0.7.1).


Original announcement

I'm pleased to announce another release of Ruby/Informix, a Ruby
library for connecting to IBM Informix.

In this release you'll find support for the INTERVAL data type, more
handy methods and more and better documentation with examples, along
with a new web site.

Web site: http://ruby-informix.rubyforge.org/
Documentation: http://ruby-informix.rubyforge.org/doc/

Some examples:

Connecting to a database:

    db = Informix.connect('stores')


Inserting records

    stmt = db.prepare('insert into state values(?, ?)')
    stmt.execute('CA', 'California')


Iterating over a table using a hash (shortcut):

    db.each_hash('select * from customers') do |cust|
      puts "#{cust['firstname']} #{cust['lastname']}"
    end



Changelog follows:

0.7.0  03/31/2008
------------------
New features:
    * Experimental support for the INTERVAL data type:
        - year_to_month, day_to_fraction, from_months and from_seconds class
          methods for creating an Interval object
        - +@ and -@ unary operators
        - +, * and / operations available with Integer, Rational, Date, Time
          and DateTime objects
        - methods for returning the respective fields of an Interval object
          individually (years, months, days, hours, minutes and seconds)
        - to_a method for returning the fields of an Interval object as an array
        - methods for converting the Interval object to the given unit, where
          apply (to_years, to_months, to_days, to_hours, to_minutes and
          to_seconds)
        - to_s method for displaying an Interval object as an string according
          to ANSI SQL standards
        - includes Comparable
    * Database#version returns a struct with version information of the database
      server.
    * Database#each and Database#each_hash shortcut methods for declaring and
      opening a cursor in a single step.
      Contributed by Reid Morrison 

    * Database#execute is not an alias for Database#immediate any more.
      It has become a shortcut for preparing and executing a statement in a
      single step.
    * SequentialCursor includes Enumerable
    * Ruby 1.9 compatible
    * More and better documentation

Bugs fixed:
    * The documentation for class Error was not being recognized by rdoc

Remarks:
    * Database.new deprecated in favor of Database.open
    * Database#do was removed
    * A lot of C code has been reimplemented in Ruby
    * Modules and classes have been reorganized
    * Database#execute still behaves the same, except that it can also accept
      input parameters and return at most one record. Database#immediate is
      more efficient though.
-- 
Gerardo Santana

Is there any way to register a datablade without using Blade Manager (also know as blademgr)?

There isn't a great deal of need for this on UNIX since the blademgr command line utility is distributed as part of IDS in the bin directory - though having said that it would be handy to be able to perform as many administrative functions as possible remotely via SQL.

On Windows there is no command line blademgr utility; the only option is to install the DBDK (DataBlade Developer's Kit) and use the GUI Blade Manager utility in order to register a blade with a database. That seems like a lot of effort to go to just to register a blade, especially since IDS 11.10 comes with some new bundled blades like Binary UDT, Basic Text Search and Node.

There is an alternative to calling blademgr, and that is to call the underlying SYSBldPrepare() and SYSBldRegister() SQL functions to register a blade manually. As far as I know this method was first published externally by Jochen Platzer in this comp.databases.informix thread in 2004: datablade registration in ANSI database

For example here is how I would register the Binary UDT datablade with a database called "mydb":

DATABASE mydb;
EXECUTE FUNCTION SYSBldPrepare('ifxmngr', 'sysblderrorlog'); 
EXECUTE FUNCTION SYSBldRegister('binaryudt.1.0','sysblderrorlog'); 

Now suppose I wanted to automate this process on Windows and create a batch file that registers an arbitrary blade to a database. Something like this would work for simple blades (i.e. blades that do not depend on other blades):

bladereg.bat:

  echo EXECUTE FUNCTION SYSBldPrepare('ifxmngr', 'sysblderrorlog'); ^
  EXECUTE FUNCTION SYSBldRegister('%2','sysblderrorlog') | dbaccess %1

So to register the Binary UDT blade to a database "mydb" using this batch file you would call: bladereg mydb binaryudt.1.0.

See Jochen's original post for instructions on manually registering blades that depend on other blades (like the Excalibur Text Blade for example). In summary, each dependent blade should be registered with SYSBldRegister in the correct order.

Is this method a supported way of registering a blade? Well, since it is not officially documented (AFAIK), it could change at some point, so if you have trouble with it, it would be advisable to try reproducing the problem with blademgr first.

I love meaningless internet popularity contests where you never really know whether the winning object of affection has the popular vote or the most orchestrated get-the-votes-out campaign. This year IBM Informix Dynamic Server 11 (IDS 11) is in the finals of the CI Product of the Year: Databases, along with Oracle Database 11g, Microsoft SQL Server, MaxDB by MySQL and CodeGear's InterBase 2007 Server Edition. So here's your chance to vote, and take part in direct democracy at its best, where sovereignty is held by the assembly of all citizens who choose to participate, or the most fanatical pressure group, depending on one's perspective.

Congratulations to Bell Micro who are the "2007 Top Performing Value Add Distributor - IBM Software" as announced in this press release today: BELL MICRO NAMED IBM’S TOP PERFORMING VALUE ADD DISTRIBUTOR FOR SOFTWARE.

From the press release:

"Bell Micro has been awarded as our best OEM Distributor in North East Europe for 2007,” said Aurore Achour, VAD Sales Manager, IBM Software Group - North East Europe. “The company, whose core offering is Informix OEM, has managed to grow both new business and maintenance OEM revenue in 2007, leading sales for North East Europe. More importantly, Bell Micro has demonstrated great value with a high number of technical certifications and skills, plus 24x7 support for both end-users and ISVs. Not only have they highly performed in OEM sales but they have also more than demonstrated their skills as Value Added OEM distributor.”

Bell Micro develop applications based on Informix technology and use these to help ISV's develop business solutions with lower hardware costs. They are also the creators of the popular Informix City website.

There have been some major problems with IDS running in Windows domains, and on domain controllers lately, and now is a good time to summarize those problems and what to do about them. I should start with a confession - we were not doing enough testing of Windows domain installation scenarios on IDS which allowed problems to creep in. As part of this investigation several new QA tests were introduced to prevent problems like this recurring.

Below is a summary of known problems with their APAR numbers which you can reference when communicating with Tech Support. APAR numbers are the defect reference numbers that begin with ICxxxx. These are only generated when a customer case is attached to a problem. If a product defect was identified and fixed internally the reference I quote starts with "idsdb" and it will be equally meaningful to tech support.

All of these problems are fixed in the latest PID (Post Interim Drop) releases of the IDS 10 and 11 codelines: 10.00.TC8W1 and 11.10.TC2W2. If necessary call tech support to discuss the status of individual defects in more detail and whether it is appropriate to receive a patch.

IC55604 WINDOWS: IDS DOMAIN INSTALL IS NOT WORKING

If the Install in Domain option is checked and if the informix 
user does not exist on the PDC/BDC, then an error message like 
the following appears:					       
Installation Message					       
---------------------------				       
The user DOMAIN\informix does not exist. Please specify an     
existing user name.					       
---------------------------				       

IC55601 WINDOWS: IDS CANNOT BE INSTALLED INTO A MICROSOFT WINDOWS DOMAIN

In a new installed and configured MS Windows domain of Windows 
Server 2003 Enterprise					       
an IDS version 10 cannot be installed.			       
After the installation was started, a new user has to be       
created during the installation process.		       
When the password for the new user 'informix' was inserted,    
and the 'ok' button was used				       
an error message box like this appears: 		       
---------------------------				       
Installation Message					       
---------------------------				       
The user DOMAINNAME\informix does not exist. Please specify an 
existing user name.					       
---------------------------				       
OK							       
---------------------------				       
Normally the information for the new user 'informix' should be 
collected.						       
Therefore this user cannot exist on this system at this time.

IC55169 WINDOWS DOMAIN USERS ARE ONLY ACCEPTED AS DOMAIN\USERNAME NOT USERNAME

You have a Windows 2000 domain style user, and a client        
computer that belongs to that domain.			       
Starting from IDS 10.00.TC6  when authenticating domain users, 
the only format accepted is				       
   DOMAINNAME\username					       
while							       
    username						       
always returns error -951.
(also reported as idsdb00153303 for IDS 11.10.xC2)

Defects which only occur in IDS 11.10

IC55490 WINDOWS: IDS 11 INSTALL ON DOMAIN CONTROLLER FAILS WITH NOROLESEP_CREATEINFXUSER 3

When IDS 11 is installed on a Windows Domain Controller on     
which the informix user does not exist the installation fails  
with the following error:				       
norolesep_createinfxuser 3 - Could not add the specified user  
to the group.						       
Although the Informix user and the Informix-Admin group get    
created, the Informix user is not added to the Informix-Admin  
group and not granted the required privileges to start IDS as a
service.						       
This problem is specific to the IDS 11.

idsdb00134811 ONINIT FAILS WITH IXDBSA GROUP ERROR AFTER INSTALLING ON WINDOWS PRIMARY DOMAIN CONTROLLER

It seems like every time Microsoft release a new service pack they introduce a DLL load address regression with it. For a little background on DLL load addresses and the problems they can cause for applications such as IDS refer to two earlier blog entries: part 1, part 2.

This time the culprit is the home networking configuration manager DLL hnetcfg.dll. In IDS installations with large shared memory configurations on Windows Server 2003 SP2 oncheck can fail with an error like this:

shmat: [22]: operating system error
MapViewOfFileEx: w32ec=487 at nt_shm.c:663
Unable to attach to shared memory.
Invalid argument

The default load address of hnetcfg.dll changed from 0x7F010000 in Windows Server 2003 SP1 (with KB913409 hotfix) back to an address space fragmenting 0x5F270000 in SP2.

To resolve the new problem, ask Microsoft for a fix of KB 948656.

ps. Thanks to Tim Steele in Australian Advanced Support for keeping track of these defects and chasing Microsoft.

On second thoughts don't learn along with me. I have the attention span of gnat. Learn along with Kate Dawson, who recently joined the Data Server Education Team in Toronto and is documenting her experiences in a new blog called Learn along with me: Data Server training, certification, and education resources. In her current post she spotlights the free certifications available at upcoming data server conferences such as the IIUG conference in April, along with available self-study tutorials, certification guides, and self-assessment exams.

Welcome to IBM Kate.

Informix security architect Jonathan Leffler is giving a webcast on Tuesday March 4th 2008 at 12PM Eastern (5PM UTC) entitled Securing Your Data: Choosing the Right Platform.

There's no doubt that information security, especially that of your database, is of ever increasing importance. Just pick up a newspaper or turn on the news and you'll hear about how yet another large company or organization has put thousands or millions of their constituents' private information at risk - and paid the penalties including loss of confidence.

Attendees will learn:

• The impact of the latest government regulations on database choices
• What even small companies need to know about database security
• How to be best prepared for audits or regulators
• How to protect your data - at rest and while 'on the wire'

The flyer for this webcast and registration details can be found here.


Draft Redbook: Security and Compliance Solutions for IDS

The draft of new IBM Redbooks publication Security and Compliance Solutions for IBM Informix Dynamic Server is now available at the IBM Redbooks web site: http://www.redbooks.ibm.com/redpieces/abstracts/sg247556.html. This Redbook provides an overview of security features in IDS and has the following contents:

Chapter 1. Technology overview
Chapter 2. Role separation
Chapter 3. Auditing
Chapter 4. Securing data with SQL
Chapter 5. Client-server communication
Chapter 6. Server-server communication
Chapter 7. Security issues with Backup and Restore
Appendix A. Audit event mnemonics
Appendix B. PAM API and macros

So take a look if you'd like to preview this book prior to final publication and perhaps provide feedback.

By the time I get around to joining a social network you can be sure it stopped being cool a long time ago. I created a MySpace profile when the only growing demographic was "sad bloggers with nothing better to do". Yesterday I finally joined FaceBook, so if you value your credibility get out now.

If you ignored this warning and still have have an account, you can become a fan of Informix Dynamic Server. I'm not sure about the concept being a fan of a product from a philosophical perspective, but am willing to make an exception for IDS, what with it being a Kult-Datenbank and all. The fan page has some interesting announcements and videos, though no photos as yet - I guess it's just a matter of time until someone adds the race car..