triBLOG

rfb: Marrying MSBuild and PowerShell

Frank Quednau — Sat, 13 Mar 2010 18:44:00 GMT

The rfb story moves on. Writing MSBuild without XML isn’t enough, it’s time to tap into the power of PowerShell and make it instantly accessible to your MSBuild script:

rfb: Marrying MSBuild and PowerShell

SharePoint 2010 / BCS / Assoziierungen können nur über die ID aufgelöst werden (und nicht über Text)

Stefan Frutiger — Thu, 11 Mar 2010 15:40:00 GMT

In SharePoint sind wir es seit längerem gewohnt Listen mit Lookup Spalten miteinander zu verknüpfen. Beim Erstellen einer Lookup Spalte kann man jeweils definieren, über welche Id die Verknüpfung angelegt wird und welches Feld zur Anzeige verwendet werden soll.

Mit SharePoint 2010 und dem neuen Konzept der External Content Types habe ich versucht diese Lookups resp. Assoziierungen zwischen Externen Listen nachzubilden.

Fazit:

Technisch gesehen funktionieren die Assoziierungen. Aber: man kann keine Anzeige-Felder definieren; Es werden immer die Id Werte angezeigt.

Beispiel:

In einer SQL Server Datenbank sind zwei Tabellen vorhanden: Event und EventYear. Zwischen den Tabellen besteht ein Foreign Key, um den Event in der EventYear Tabelle benutzen zu können.

Diese beiden Tabellen wurden in SharePoint Designer 2010 als External Content Types registriert. Zusätzlich wurde eine Assoziierung zwischen den External Content Types erstellt, um den Foreign Key (resp. den Lookup) in SharePoint nachzubilden.

Danach habe ich die External Lists basierend auf den Content Types erstellt. wenn ich im Browser in die Event Liste anschaue, sehe ich folgendes:

So weit alles in Ordnung. Der Event (also das ListItem) kann im Browser angezeigt und modifiziert werden. Toll!

Bei der EventYear Liste sieht es hingegen anders aus:

Die ListItems können zwar editiert werden; aber der Event wird als Integer Id dargestellt und nicht als Text.

Anscheinend soll es auch in Zukunft nicht möglich, dass sich Assoziierungen zwischen External Content Types gleich verhalten wie die Lookup Spalten innerhalb von SharePoint Listen.

Externe Referenzen:

http://social.technet.microsoft.com/Forums/en/sharepoint2010customization/thread/c382b04f-b55e-4c65-b9e8-b295b48eebd0

Inserts Experiencing an Increasing CPU Consumption

Striving for Optimal Performance — Mon, 08 Mar 2010 21:23:42 GMT

Last week I had to analyze a strange performance problem. Since the cause/solution was somehow surprising, at least for me, I thought to share it with you. Let me start by quickly describing the setup and what was done to reproduce the problem: Database version: Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 (64-bit) Operating system: Solaris 10 (SPARC) To [...]...(read more)

A few things that have piled up

Frank Quednau — Mon, 08 Mar 2010 09:14:00 GMT

Well, unfortunately community server cannot read my shiny new self-made rss feed (well not quite, it is based on the relevant WCF classes). Even so, I’m not that sad, since I had the glorious (for a Monday) idea to update this blog with Windows live Writer, which makes this reeal’ easy.

Since the introduction of my ASP.NET MVC –based website I’ve tried to gather opinions on getting an Open Source distribution system going for .NET . Reception so far has been…none. Shouldn’t be too surprised since there isn’t any code, and some people seem to think that “which has not been coded, does not exist”. There’s a bit of truth in that.

I first had to get something else out of the door, the first version of the rfb. Slogan: MsBuild, but without XML. It’s a first version, and in some ways it is a proof of concept due to current limitations, but it can already run the build script I use for my Rf.Site project.

Next will be a proof-of-concept of the aforementioned and then trying to get to talk to those Open Source projects that stand at the beginning (or end?) of the foodchain: Things like Moq, Castle Dynamic Proxy, LinFu, log4net. I’ll keep you informed how it goes.

JUGS Event über Patterns und Best Practices mit GWT

paw — Fri, 05 Mar 2010 08:24:00 GMT

Am 2. März besuchte ich einen interessanten Vortrag über GWT. Heiko Braun, ein Senior Software Engineer bei JBoss, hat über seine Erfahrungen mit GWT berichtet.

Momentan ist er mit der Entwicklung des SOA Tooling für JBoss beschäftigt. Das Tooling muss sehr Modulare Eigenschaften haben, was bedeutet, dass auf der einen Runtime (z.B. Tomcat) gewisse Tools zur Verfügung stehen und auf anderen einer anderen (z.B. Jetty) nicht. Diese Modularität bietet GWT nicht out of the box. Des weiteren traten die bekannten Probleme bei Rich Client Software, wie Koplung zwischen Komponenten und Services, auf. Als Antwort auf die Probleme haben sie die einzelnen Lösungen aus dem SOA Tooling herausgelöst und als eigenes JBoss Projekt der GWT Community zur Verfügung gestellt: Project Errai

Das Projekt Errai ist als erste Version verfügbar und bietet drei Hauptkomponenten an:

Die Basis des Projekts bildet ein Message Bus. Darauf wurde ein Workspace Framework gebaut, das den Ansatz eines Eclipse RCP bzw. Netbeans Plattform verfolgt. Als Goody werden alle speziellen Widgets, die während der Implementation des SOA Tooling entstanden sind, ebenfalls zur Verfügung gestellt. Die Widgets sind jedoch nicht als Zentralen Bestandteil zu betrachten. Die Abhängigkeiten der einzelnen Komponenten kann aus dem Diagramm entnommen werden (der Workspace benötigt also den Message Bus).

Meiner Meinung nach ist der Message Bus ein sehr interessantes Teil, das eine wichtige Anforderung im Bereich Rich Client Programmierung Abbildet. Da die Interaktion der einzelnen Modularen Teile untereinander meist ein zentraler Punkt (auch Knackpunkt ) darstellt, bietet dieser Message Bus interessante Möglichkeiten:

client-client across server (chat server)
client-client w/o server (inter component)
client-server (client send)
server-client (server push)

Natürlich sind das keine neuen Konzepte, man muss normalerweise jedoch mehrere Technologien unter einen Hut stecken, damit man diese Funktionalität erhält. Zu guterletzt hat der Bus noch ein sehr schönes API im Fluent Interface Style nach Martin Fowler und konfiguration durch Annotationen.

Das Workspace Framework ist auch nicht zu missachten. Es ist in Eclipse RCP Manier aufgebaut, wenn auch etwas reduziert:

Um ein einzelnes Tool im Workspace zu registrieren genügt eine Annotation:

Das Projekt hat noch viele andere Features anzubieten. Am besten geht ihr an einen Vortrag darüber, oder schaut euch die Projektpage bzw. Doku genauer an. Meine Neugier hat das Projekt auf jeden Fall geweckt.

Die Folien sind bei JUGS / Vorträge zu finden. Gewisse Bilder sind aus diesem Vortrag und aus der Projekt Errai Page übernommen.

btw: Habt ihr euch schon mal gefragt, was JUGS übersetzt bedeutet? -> dict.leo.org/jugs

Project Management with TFS 2010

Marc Cortesi — Thu, 04 Mar 2010 15:16:04 GMT

Compared to earlier versions, TFS 2010 offers a number of improvements regarding project management. Requirements of different stakeholders can now be captured and subdivided into so-called scenarios. A scenario in turn consists of the well-known work items. This structure makes a top-down planning approach much more native.

Also new is the enhanced integration in MS Project and Excel. So far, exports and imports of work items from and to Project were possible, but hierarchies and adjustments inside the Project files got lost. Furthermore a connection to MS Project Server was only possible with third-party tools. Microsoft bridged this gap and made the integration with the well-established tools complete.

The new principle of user definable link types makes it possible to create successor/predecessor relationships. In the earlier versions, work items only could be linked to each other in a peer-to-peer manner. With those links it wasn’t possible to create a tree structure or view dependency of one work item over other. The process templates already provide different link types, but even more interesting is the fact that TFS 2010 offers a framework for creating custom link types.

Tracing VPD Predicates

Striving for Optimal Performance — Sun, 28 Feb 2010 11:30:16 GMT

Even though a number of articles and blog posts have already been written on this topic (e.g. on Pete Finnigan’s site I found references dating back from 2003), from time to time I’m still asked “How to trace predicates generated by VPD?”. Hence, here’s yet another blog post about this topic… Let’s setup the scene before [...]...(read more)

BASTA! Spring 2010 ::: WPF 4.0 – Die Neuerungen im Überblick

.NET Rocker » TrivadisContent — Thu, 25 Feb 2010 14:29:33 GMT

Danke für die zahlreiche Teilnahme an meiner Session zu den Neuerungen in WPF 4.0. Ich hoffe, es hat allen Spass gemacht und jeder konnte etwas mitnehmen. Folgend die versprochenen Slides und Demos: Slides Demos Happy Coding, Thomas...(read more)

BASTA! Spring 2010 – Model-View-ViewModel

.NET Rocker » TrivadisContent — Wed, 24 Feb 2010 17:36:55 GMT

Hallo liebe Freunde der BASTA!, VB.NET'ler und Teilnehmer an meiner Session zum Model-View-ViewModel-Pattern. Folgend die versprochenen Slides und die Beispielanwendung zum Download. Slides MVVM.BeispielApp Viel Spass damit, Thomas...(read more)

From Visual Studio 2010 Release Candidate back to Beta 2

.NET Rocker » TrivadisContent — Mon, 22 Feb 2010 14:44:50 GMT

If you’ve played around with Visual Studio 2010 Release Candidate (RC), you sure have noticed that it’s pretty fast. E.g. the WPF- and Silverlight-Designers come up quickly and much faster that in Visual Studio 2008. But for now there are some reasons to wait before installing Visual Studio 2010 RC: Silverlight 4 Beta is not supported. Silverlight [...]...(read more)

Be careful with Default-Values of Dependency-Properties if you’re using Reference-Types.

.NET Rocker » TrivadisContent — Thu, 11 Feb 2010 13:35:50 GMT

I’ve justed finished a small PieMenu for a WPF-Application for one of our customers in switzerland. The PieMenu is displayed below with a black-red-style. The Style, colors and items can be changed individually. The colors and icons etc. in the PieMenu below are just the ones I’ve created for testing purposes. While developing this PieMenu [...]...(read more)

Visual Studio 2010 RC and Silverlight 4 Beta

.NET Rocker » TrivadisContent — Tue, 09 Feb 2010 15:39:45 GMT

Yesterday Visual Studio 2010 RC was released to MSDN Subscribers (find the link here), tomorrow it’s available for download for everyone. The performance is great as far as I can say by using it for at least one day. But it lacks on support for Silverlight 4 Beta. If you’re developing Silverlight 4 applications, it is [...]...(read more)

Visual Studio 2010 Release Candidate (RC) available next week

.NET Rocker » TrivadisContent — Sat, 06 Feb 2010 13:10:55 GMT

The launch date for Visual Studio 2010 RTM is the 12th of April. This month there’ll be a Release Candidate (RC) and Scott Guthrie mentioned on his blog in a post about ASP.NET MVC 2 RC that it will be available soon. Now everybody speculates about the date and what “soon” means. Days or weeks? The [...]...(read more)

Ausserplanmässiger Security-Fix für WebLogic

Sven's Technik-Blog » TrivadisContent — Fri, 05 Feb 2010 06:57:41 GMT

Oracle hat gestern einen Security-Fix ausserhalb der quartalsweise erscheinenden CPUs veröffentlicht. Dieser schliesst eine Lücke im im WebLogic Node Manager (CVE-2010-0073), welche remote ohne Authentifizierung ausgeführt werden kann. Der CVSS-Basescore ist 10 – es kann also zum kompletten Verlust der Integrität, Vertraulichkeit und Verfügbarkeit führen! Deshalb empfiehlt Oracle den Patch sofort einzuspielen. Mehr Informationen und der Link zum [...]...(read more)

The Art of Application Performance Testing

Striving for Optimal Performance — Sun, 31 Jan 2010 13:52:38 GMT

If you are looking for a book that describes how to approach and carry out a performance testing project, don’t look further. Get a copy of The Art of Application Performance Testing: Help for Programmers and Quality Assurance (O’Reilly, 2009). In it Ian Molyneaux manages to cover, in 160 pages, a crucial topic that [...]...(read more)

The Last Drupal Post

realfiction - TrivadisContent — Sat, 30 Jan 2010 17:32:01 GMT

The Domain transfer of realfiction.net is almost through, which means that I say goodbye to the Drupal CMS and hello to Rf.Sites, my own codebase to publish stuff. If availability is gone at some point it means that the new nameservers are being propagated through the webs. Redirects should be in place for the most important URLs, however, this one post will not be migrated anymore, so enjoy while it's there :)...(read more)

Does the Query Optimizer Cost PX Distribution Methods?

Striving for Optimal Performance — Tue, 26 Jan 2010 11:55:15 GMT

The short answer to this question is “yes”, it does. Unfortunately, the distribution costs are not externalized through the execution plans and, as a result, this limitation (yes, it is really a limitation in the current implementation, not a bug) confuses everyone that carefully look at the information provided in an execution plan of a [...]...(read more)

Speaking at BASTA! Spring about WPF 4.0 and MVVM

.NET Rocker » TrivadisContent — Sun, 24 Jan 2010 16:20:47 GMT

The BASTA! (=Basic Days => in German Basis Tage) is “the” Conference in Germany about .NET. The BASTA! Spring is from 22nd to 26th February. I’ve two sessions there: one Wednesday (24. February) and one on Thursday (25th February). I’ll talk about the new features in WPF 4.0 and about the Model-View-ViewModel-Pattern (MVVM). Find more (German) [...]...(read more)

Using the event API to publish an event to the Event Delivery Network (EDN) – the Spring way

Enjoy IT - SOA, Java, Event-Driven Computing and Integration » Trivadis — Wed, 13 Jan 2010 00:59:13 GMT

The Event Delivery Network (EDN) in Oracle SOA Suite 11g provides a declarative way to use a publish/subscribe model to generate and consume business events without worrying about the underlying message infrastructure. Events can be published / subscribed from a variety of programming environments such as Java, PL/SQL, SOA Composites, and ADF-BC applications In his blog [...]...(read more)

Edition-Based Redefinition

Striving for Optimal Performance — Tue, 12 Jan 2010 18:42:58 GMT

Upgrading critical applications can be very difficult. One of the main problems is that for reasons of availability, long downtimes cannot be periodically scheduled. Therefore, for such applications, it is desirable to implement online upgrades. This requires that the application in question, as well as any software used by the application (e.g. the database engine) [...]...(read more)

Oracle optimizer system statistics are critical !

Yann Neuhaus — Mon, 11 Jan 2010 20:04:00 GMT

The goal of this post is not to describe a tuning session; as reminder, this blog has been created to allow information exchange around Oracle High Available infrastructures J. However this week I really wondered how "corrupted" Oracle SYSTEM Optimizer statistics (not object statistics) could have a HUGE impact on the database performances.

Indeed, a customer of mine migrated from Oracle 9i on TruCluster to Oracle 11.1 on Linux Clusterware. Nothing special. He performed several import tests and application tests. During the validation phase, everything worked fine. In the week-end of the year switch (2009 to 2010) the production database was migrated for the real "go-live". The migration worked fine but on Monday the performances were quite poor. Very strange because the week before, during the validation tests, the application worked fine.

The customer didn't change anything between production and validation, he just collected the system statistics on the morning of the go-live.

After a deeper analysis we observed with the developers, that none of the explain plan corresponded to the expectations. The selective index was ignored, only the index on the columns referenced in the "order by" clause (but not efficient) were used.

As you may also experienced in projects, a huge SQL troubleshoot session has been started, analysing in depth all the plans, checking the optimizer parameters, and so on....

Let's take this simple SQL as example, the function based index on "SUBSTR(NLS_LOWER("DATAAREAID"),1,4), SUBSTR(NLS_LOWER("SALESID"),1,20)" is used even if the index on "Substr(Nls_lower(customerref),1,60)" would be much more efficient.

SELECT /*+ FIRST_ROWS (100)*/ a.salesid,
a.salesname,
a.reservation,
a.custaccount,
a.invoiceaccount,
a.deliverydate,
a.deliveryaddress,
a.url,
...............a.recid,
a.del_correctivereason,
a.cibzvkgreetingtext
FROM     salestable a
WHERE    ((Substr(Nls_lower(dataareaid),1,4) = Nls_lower(:in1))
          AND ((NOT ((returnstatus = :in2))
                AND NOT ((returnstatus = :in3)))
               AND (Substr(Nls_lower(customerref),1,60)LIKE Nls_lower(:in4) ESCAPE N'\')))
ORDER BY Substr(Nls_lower(a.dataareaid),1,4) DESC,
         Substr(Nls_lower(a.salesid),1,20) DESC

Index I_366SALESIDX, index Columns ("bad index") :

Column Name Data Type
SUBSTR(NLS_LOWER("DATAAREAID"),1,4) COLUMN EXPR.
SUBSTR(NLS_LOWER("SALESID"),1,20) COLUMN EXPR.

Index I_366CIBZALCUSTOMREFIDX, index Columns ("good index") :

Column Name                             Data Type
SUBSTR(NLS_LOWER("DATAAREAID"),1,4)     COLUMN EXPR.
SUBSTR(NLS_LOWER("CUSTOMERREF"),1,60)   COLUMN EXPR.

Having a look at the in explain plan, we see that the not effective index on SUBSTR(NLS_LOWER("SALESID"),1,20) is used.

While using values for the bind variables we obtain the following SQL execution statistics :

------------------------------------------------------------------------------
| Id | Operation                    | Name          | Rows | Bytes | Cost |
------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |               |     1 | 1435 | 3088K|
|   1 | SORT ORDER BY               |               |     1 | 1435 | 3088K|
|   2 |   TABLE ACCESS BY INDEX ROWID| SALESTABLE    |     1 | 1435 | 3088K|
|   3 |    INDEX RANGE SCAN          | I_366SALESIDX | 3865K|       | 31656 |
------------------------------------------------------------------------------

Statistics
----------------------------------------------------------
        316 recursive calls
          4 db block gets
     447118 consistent gets
     234802 physical reads
          0 redo size
      15011 bytes sent via SQL*Net to client
       3044 bytes received via SQL*Net from client
          2 SQL*Net roundtrips to/from client
          0 sorts (memory)
          0 sorts (disk)
          0 rows processed

When observing the Optimizer System statistics, we discovered the following issues :

select * from aux_stats$

SNAME           PNAME           PVAL1      PVAL2
--------------- --------------- ---------- ----------------
SYSSTATS_INFO   STATUS                     COMPLETED
SYSSTATS_INFO   DSTART                     01-04-2010 09:45
SYSSTATS_INFO   DSTOP                      01-04-2010 10:45
SYSSTATS_INFO   FLAGS                    0
SYSSTATS_MAIN   CPUSPEEDNW         542,903
SYSSTATS_MAIN   IOSEEKTIM            6,958
SYSSTATS_MAIN   IOTFRSPEED       21836,194
SYSSTATS_MAIN   SREADTIM            28,136 <= ! Too high !
SYSSTATS_MAIN   MREADTIM            66,433 <= ! Too high !
SYSSTATS_MAIN   CPUSPEED               611
SYSSTATS_MAIN   MBRC                    12
SYSSTATS_MAIN   MAXTHR           386176000
SYSSTATS_MAIN   SLAVETHR           3713024

SREADTIM, MREADTIM had HUGE values.

Single Read Time has 28 ms of wait time ! This is for such a system almost impossible since it enjoys quite fast hardware (last HP EVA technology, last HP server, HBA for connection to disks, aso ...). While analysing the average I/O waits with Grid Control we observed between 1 and 4 ms wait time.

The customer remembered that these system statistics were calculated during quite heavy load.

We started again the system statistics calculation during the lunch break (during 30 minutes with normal/low load) and observed the results :

exec dbms_stats.gather_system_stats
(gathering_mode => 'interval',interval => 30);

SQL> select * from aux_stats$;

SNAME           PNAME           PVAL1      PVAL2
--------------- --------------- ---------- -----------------
SYSSTATS_INFO   STATUS                     COMPLETED
SYSSTATS_INFO   DSTART                     01-05-2010 12:25
SYSSTATS_INFO   DSTOP                      01-05-2010 12:55
SYSSTATS_INFO   FLAGS                    0
SYSSTATS_MAIN   CPUSPEEDNW         542.903
SYSSTATS_MAIN   IOSEEKTIM            6.958
SYSSTATS_MAIN   IOTFRSPEED       21836.194
SYSSTATS_MAIN   SREADTIM             6.565
SYSSTATS_MAIN   MREADTIM             7.939
SYSSTATS_MAIN   CPUSPEED               825
SYSSTATS_MAIN   MBRC                    14
SYSSTATS_MAIN MAXTHR            10621952
SYSSTATS_MAIN   SLAVETHR              1024

Now the single block read time (SREADTIM) is 6,5 ms, the maxthr (means maximum I/O throughput in bytes/s) slowed down from 386 MB/s to 10MB/s. This was due to the lower activity on the server during system statistics collection.

The optimizer directly integrated these changes, and the plan now uses the right/selective index : I_366CIBZALCUSTOMREFIDX

While using values for the bind variables we obtain the following execution statistics :

----------------------------------------------------------------------------------------
| Id | Operation                    | Name                    | Rows | Bytes | Cost |
----------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |                         |     1 | 1435 | 2931 |
|   1 | SORT ORDER BY               |                         |     1 | 1435 | 2931 |
|   2 |   TABLE ACCESS BY INDEX ROWID| SALESTABLE              |     1 | 1435 | 2930 |
|   3 |    INDEX RANGE SCAN          | I_366CIBZALCUSTOMREFIDX | 3805 |       |    20 |
----------------------------------------------------------------------------------------

          0 recursive calls
          0 db block gets
       3933 consistent gets
       2867 physical reads
          0 redo size
      14982 bytes sent via SQL*Net to client
       3014 bytes received via SQL*Net from client
          2 SQL*Net roundtrips to/from client
          1 sorts (memory)
          0 sorts (disk)
          0 rows processed

Costs dropped down from 3088K to 2931.

As a conclusion we could state that we solved the performance problem just by choosing the right interval to collect the system statistics. The important of these system statistics is well described in the book of my friend Chris (http://antognini.ch/top/).

Join Elimination

Striving for Optimal Performance — Mon, 11 Jan 2010 17:50:30 GMT

In some specific situations the query optimizer is able to completely avoid executing a join even if a SQL statement explicitly calls for it. Two are the cases currently covered by this optimization technique, which is called join elimination. The first one was introduced in Oracle Database 10g Release 2, the second one in Oracle [...]...(read more)

Using the new Object/XML Mapping Support of Spring 3.0 with JiBX and Maven

Enjoy IT - SOA, Java, Event-Driven Computing and Integration » Trivadis — Wed, 06 Jan 2010 19:17:28 GMT

The Object to XML mapping functionality (OXM) has been moved from Spring Web Services project to the core Spring Framework with version 3.0 of Spring. Object/XML Mapping, or O/X mapping for short, is the act of converting an XML document to and from an object. This conversion process is also known as XML Marshalling, XML Serialization [...]...(read more)

Solving javax.naming.NameNotFoundException: Unable to resolve ‘jdbc.SOAAppUserDataSource’ with Oracle SOA Suite 11g Adapter Services

Enjoy IT - SOA, Java, Event-Driven Computing and Integration » Trivadis — Fri, 01 Jan 2010 14:22:35 GMT

I run into this rather “stupid error” a couple of times already. It’s very easy to solve, however it’s also very easy to do it wrong when configuring Oracle SOA Suite 11g Adapter services through the WebLogic Console. I hope this blog entry will help new users when running into this problem: You get the following error [...]...(read more)

Posting Source Code with Microsoft Live Writer and WordPress

Enjoy IT - SOA, Java, Event-Driven Computing and Integration » Trivadis — Sun, 27 Dec 2009 19:32:53 GMT

Today I searched for a solution to post source code using Microsoft Live Writer to my blog hosted on WordPress.com. WordPress itself has an integrated Syntax Highlighter supporting most popular languages. Details of how to use it can be found here. But how do you use the WordPress Syntax Highlighter when using Microsoft’s Live Writer (which I [...]...(read more)