Today is the first day of the DB2 10 for Linux, Unix and Windows bootcamp at IBM Hursley and so far it has been informative. A special thank you has to go out to Triton Consulting and Iqbal Goralwalla (@iqbalgoralwalla) Head of mid-range at Triton for getting IBM to put one of these on in the UK. With a thank you also going to IBM for doing all it for free!
DB2 10 for Linux, UNIX and Windows Bootcamp – Day 1 – What have we done – my prospective
First gem that the course instructor let onto and I did not spot anywhere so far (twitter, channel db2 etc) is that there is a fixpack for 10 already, that fixes some early bugs that have been reported.
DB2 Features Spotlight
Basically going through all the details and features that make DB2 V10 a good proposition, wetting you appetite like any good sales pitch should do. The are promising improving the pillars of DB2, low cost operation, ease of development and reliability. In low cost operation they are offering things like adaptive compression and multi temperature data management. Multi temperature data management is the most interesting here as it raises interesting questions around if you have a SAN that is clever enough to realise the hot blocks of data is it better to let the SAN handle hot data or is it better to set up storage groups and define hot storage groups.
DB2 Fundamentals
This section before lunch was split into two main parts, one covering a lot more than the other. The first was a product overview and the second was the fundamentals. In the product overview I was surprised to see that HP-UX was still a supported platform and that there was no mention of the mobile edition that has been tweeted about a lot by the @DB2_IBM account recently. Also raising some questions over the INGEST utility being available in which edition, which according to the matrix which is in the notes then it is only in Advanced Enterprise Server Edition which is not the version that I work with, and put a spanner in the works for our future plans.
The fundamentals section is little changed from DB2 V9.7 and its fundamentals. There are a few changes to commands, and extensions to others. You can now get your db2diag log to rotate based on size, which is useful, but I would have preferred the splitting to happen based on time, like every 24 hours.
Bye Bye Control Centre, Hello data studio
This was an attempt to introduce the IBM Data studio that replaces Control Center. I am still reserving jugement on Data studio as it does not work very well on my Windows VM with 4 CPUs and 3GB of RAM (Mac, VMWare Fusion), but on the SUSE linux VM that was running 2 CPU’s and 2GB of RAM (Windows 7, VMware) seemed to run a LOT better for our hands on lab.
You will also need to upgrade your Data studio to 3.1.1 to exploit all the features of DB2 V10. This means I might invest some time in setting up a SUSE Linux Vm and get this working properly. We did not go through the web console, but that might be later in the course.
Storage design and Multi-Temperature Storage
This covered some of the older concepts of the bufferpools and table-spaces and how they have been spiced up with the introduction of storage groups and how these can be set up to create Multi-Temperature data and manage where it is stored. I think this will be interesting and will lead to many debates with your storage architect over which level (storage or application) will decide where data is stored.
We did a hands on lab using some of these concepts and it was quite interesting, but the small size of the SAMPLE database belittles the REBALANCING and other issues around moving storage groups and adding and removing storage areas.
My main point here would be interesting to see the differences between DB2 being set up on storage with SSD etc that are in the IBM examples VS hot block swapping on the SAN and non of the Multi-Temperature data management, and see which is more efficient and over what types of workload; but who has the time or the resource to justify doing this, (IBM?)
Basic Maintenance & Autonomic Features
Some of the concepts looked into here seemed to some as a bit of a shock to some of the delegates, but not to us, apparently needing to REBIND packages after you have carried out RUNSTATS operations was not the done thing. Covering STMM and other automatic features and there was not a lot in here that was new.
Tomorrow
Tomorrow we will do the Hand on Lab for Basic Maintenance & Autonomic Features, and carry on working through the slides and the labs, should be good.