Cloud Foundry Logging Sources Deciphered

Ever deployed a cloud foundry app and ran into errors? I did and still do. My typical reaction is to request the recent app logs and to analyse them for the root cause. The logs contain those strange (if you don't know them) codes, e.g., RTR, STG, APP, PROC, WEB or more. Here is how to decipher them and use them to your advantage.

Happy Cloud Debugging

The codes above denote the Cloud Foundry component emitting the log entry. You can find a list of those component identifiers in the Cloud Foundry documentation or in the IBM Cloud (Bluemix) documentation on Log Sources. There are many component names. I found that remembering the following helps a lot:

• RTR: This is the router which channels HTTP requests to an app. Here you find information about the request itself including client information.
• STG: The staging phase has details about the deployment and restaging of an app. They help with initial errors.
• APP: If your app writes information to stderr or stdout, then you will find it here. For some apps this could be the well-known "printf debugging".

The log entries have either an OUT if the information is from stdout or ERR if from stderr. Else there is not much to remember. Happy Debugging. If you have feedback, suggestions, or questions about this post, please reach out to me on Twitter (@data_henrik) or LinkedIn.

Puzzling: Enterprise Amnesia vs. Enterprise Intelligence (IBM Redbook Video)

A very interesting video featuring IBM Fellow Jeff Jonas has been published on the IBM Redbooks site. The title is Enterprise Amnesia vs. Enterprise Intelligence. Jeff talks about how organizations have trouble making sense out of the growing amount of data, how Big Data changes some of the existing laws ("physics").

In the video a puzzle is used as metaphor and practical example on how context information can be used to make sense of data. What was also interesting to me was the use of Ravensburger jigsaw puzzles for his experiment (watch the video!). The company Ravensburger is located in Ravensburg in direct vicinity to Friedrichshafen. Interesting to see how old fashioned stuff from the Lake Constance region helps understand leading edge IBM technology for Information Management. Still puzzled? Watch the video...

Consistency checks: Failed Groupon email

This morning I had the usual mix of emails in my private inbox: Filtered and unfiltered spams, some emails directly for me, and a set of marketing emails. When I wanted to delete an email sent by the German Groupon site, I spotted something that didn't make sense. My first reaction was to re-read the email, then to press the button to load the "remote content" (for safety and privacy reasons this is disabled by default). Now with pictures, the content still didn't make sense:

Screenshot of Groupon email

As a database guy and also lecturer my reaction was that this will be a great talking point when the topic of consistency checks and constraints is on the agenda again. But how to make sure, I can always access it and point others to it? Here we go... ;-)

Turn XML data into columnar, relational format using SQL (lots of useful links inside)

The SQL standard offers many functions, some very useful are defined in part 14 of the SQL standard, "ISO/IEC 9075-14:2011 Part 14: XML-Related Specifications". That part 14 was published 2003, so it is rather old. However, one of the gems in SQL/XML (the name for that XML addition to SQL) and in DB2, the function XMLTABLE and its capabilities seem to be hidden as a recent email to me indicates. Or is it how Google, Bing, and other are used...? Anyway, in this article I will provide pointers to some useful resources, mostly older articles, and tag them with the right buzzwords. All in the hope that it is one email less in my inbox...

On IBM's developerWorks is a two-part article giving a detailed overview about XMLTABLE. In part 1 we learn about all the ways XML data can be turned into relational format, including best practices. The 2nd part deals with advanced processing, such as splitting XML documents into smaller pieces (including parameter passing), shredding XML documents during insert processing (turn XML into tables during insert), relational views over XML data, and creating field/value pairs out of XML data (is this already NoSQL...?). Parameter passing for XML processing is also a topic of one of my older posts. XMLTABLE can also be used to extract XML data embedded into another XML document via CDATA. And last, a useful webcast that discussed XMLTABLE in DB2 for both the mainframe and the distributed platform can be found at the pureXML devotees. That group, now driven by IBM customers using pureXML, has upcoming meetings/webcasts. Last but not least, Matthias Nicola has written many articles about XMLTABLE in his blog.

Let's see what the search engines do with this article. Have a nice week...

SQL and XML Limits in DB2

How many database partitions can a DPF system have (it is 999 partitions)? How many levels deep can an XML document stored in DB2 be (125 levels)? What is the maximum size of an SQL statement I can code (2097152 bytes)?

If you have questions like that around DB2, then the page SQL and XML limits in the DB2 Information Center is the right one for you. And if you have lots of users, rest assured, the maximum number of concurrent users for a system is 64000.

The importance of correct and integrated data

You can deliver a product as fast and "good" as promised, but still loose a repeat customer. Here is a real life example.

Over the past days I ordered several books online. Usually, I use one specific company "A" (any guesses?) for two reasons. First, in Germany by law all German language books have to be sold for the list price, neither discounts nor coupons can be applied. "A" is offering free shipping on all books, regardless of their list price. Second, "A" is usually very quick in delivering the books and provides tracking information. I used this company as usual for some books.

Then, for an English book where every book seller can offer individual prices, I shopped around and found that I could save 2 EUR over at company "B". The product details showed that the book was in stock and would be sent the next day. I ordered the book in the evening and then started to wait for a confirmation email. And waited. The next morning I received 2 emails. The first was the acknowledgement of my order. The second was a status change email - now the book wasn't in stock anymore, but had a delivery time of 3 weeks. Ouch.
I checked the order status online and then the product page again. Both showed the 3 weeks for delivery. Why not cancel the order and shop somewhere else? I sent an email via contact form, tried to cancel online, and called the service hotline. On the phone they were friendly, confirmed the change to 3 weeks, wondered why it had changed over night and pointed out that it would be hard to cancel, but they would try anyway. The reason was they are working with a 3rd party fulfillment center. One day later, first the status said "trying to cancel", then it changed to "package sent", and in the afternoon the book arrived.

Company "B" delivered the product I ordered on time. However, their data isn't integrated and both customer and customer service are left to wonder about what is correct. Will I shop there again? Not this year... ;-)

Sent from my... (Information Overload)

Yesterday evening I stumbled over a funny (non-IBM!) video about information overload. IBM also has a podcast on Information Overload. Too much is too much. But how much information is ok? And what information do you need? Do you get the right information? At the right time? At the right place? Are you able to make use of that information? Do you pay too much to get the right information?

IBM calls the process around answering these questions to establish an Information Agenda. Not as funny as the video, but certainly effective.

BTW: This message was not sent from a Blackberry.