Lukas Beeler's IT Blog

Out of ideas. Sorry.

I’m currently Outlook 2007 with an Exchange 2007 server for all my business needs. But of course i also have a private domain (which this blog runs on), and where i have my private e-mail.

For a long time, i’ve used Mutt (a terminal based text e-mail client) to handle all my private mails. It worked great, especially filtering, threading, etc. were very well done. But the disadvantage are obvious. So i’ve searched for a suitable replacement. I’ve tried a variety of clients, but didn’t really like any of them.

Then i’ve tried Gmail. I was quite pleased with it, but it didn’t offer enough storage to hold all my mails. So i’ve subscribed to Google Apps Enterprise, with a single user account, and an approriate forward on my Postfix setup.

Apps Enterprise can use IMAP to migrate your mails from your old server, and tag the mails according to the folders. Worked perfectly, but took around a day (for ~250’000 mails).

So far, i’m quite pleased, even though i’m not really sure if i like Gmail’s threading style. Google talk can also relay incoming mail notifications.

I was unable to use Gmail as my primary MX, because it is impossible to add an E-Mail route (you can add it in the GUI, but it doesn’t get saved). I opened a case about this two weeks ago. No solution yet.

ExcelliPrint is an excellent tool for IPDS->PCL conversion, and i’ve written about it many times before.

But it is only as good as it’s users. A customer called me with a strange problem, related to ExcelliPrint. He told me that he can print on the System i just fine, but the print job just vanishes into thin air. The System i showed the file as being printed, but it was nowhere to be seen on the Windows machine.

I asked the customer if he changed anything recently, he told me no. I’ve used Netviewer to access the customers servers, and have a look at ExcelliPrints logfiles.

At that point, the problem was obious: ExcelliPrint complained that the printer name was invalid. I checked this against the installed printers, and of course the printer ExcelliPrint had another name. After selecting the correct printer, ExcelliPrint continued working just fine and reliable as always.

When asking the customer why he renamed printers, and didn’t tell me that he just said “I didn’t think it mattered”. So much.

Did you know that you can enable Remote Desktop/Terminal Server to use SSL?

Configuration authentication and encryption for Terminal Services

It is generally good practice to configure any machine which has Remote Desktop or Terminal Services enable to at least have an SSL certificate that can be used with RDP. It’s easy to do, and it will allow RDP to use better encryption.

This is especially important if you’re running RDP directly over the Internet (for which special care needs to taken in many more aspects), but it also makes sense to use this in local LAN.

If you don’t have any legacy clients, it also makes sense to set the accepted keystrength to “High”. This will cause all older RDP clients to fail. If you can’t risk that, you can still use “client-compatible”, and use SSL with newer clients and RDP’s builtin encryption with older clients.

Debugging Wireless LAN has always been a rather difficult item. And it was even more difficult on Windows, because you didn’t see many things that other operating systems showed you at point blank range.

But Microsoft has a commandline tool available that many people do not know about, but might make your life a lot easier if you do.

netsh wlan show networks mode="bssid" interface="Drahtlosnetzwerkverbindung"

You’ll have to replace “Drahtlosnetzwerkverbindung” with the name of your wireless adapter. Here’s a snippet of example output:

SSID 1 : dataline
Netzwerktyp : Infrastruktur
Authentifizierung : WPA-Enterprise
Verschlüsselung : TKIP
BSSID 1 : 00:19:07:90:91:00
Signal : 100%
Funktyp : 802.11a
Kanal : 116
Basisraten (MBit/s) : 6 12 24
Andere Raten (MBit/s) : 9 18 36 48 54

Nice and detailled, isn’t it? Much better than the GUI. I would strongly recommend any Windows admin to read up on netsh. It offers many possibilities and debugging option, yet this tool isn’t as well understood as it should be.

I’ve wandered across an interesting product today.

GUI5250

Looks old and dated? It isn’t.

XTendGUI wird voraussichtlich ab November 2007 verfügbar sein.

Or, in English: XTendGUI should be available starting November 2007.

I’ve spent most my week with working with reporting bugs for IBM’s new DB2 Web Query for System i.

I’ve written mostly about my negative experiences, because there wasn’t much else to be done for me. With the most basic tasks like the German language version and Query/400 import failing, we couldn’t experience all the cool things that QU supposedly offers.

While reading the appropriate Redbook and doing the exercises located within, it looks like a necessity to create some more complicated views to avoid of having to do too many custom joins in the report itself. But this entirely depends on your database structure. QU2 can extract information from your foreign keys in order to generate joins automatically, but this of course only works if you’re already using foreign keys (you should!).

As things are right now, all PMR’s i’ve talked about so far are still open and unfixed, mostly because they’re not in IBM’s direct responsibly and have to be forwarded to Information Builders. And they’re all for very basic functionality like changing the language of the product, or of it’s key features – it works properly in an U.S. English environment, but not yet in a #2939 Swiss German environment.

I will hold a small in-house demo to get word about this new product out and i hope that my coworkers have a bit of a more positive mind about the product than i do. We will see. I’ll post progress of my still-open PMRs, IBM is working on a fix, so i’ll hope it is resolved quickly.

Straight from PMR 61860,212,848 – the Query/400 import issue (details here).

For a circumvention you can try either of these:
1) Try to create the metadata with a usrprf with CCSID 37
2) Use this other console (we call it the “metadata console”) at port 11333. Click on “Add Metadata”, Data Adapters, the screen after that will look similar to port 11331

Oh, and there is more fun to be had. The overview list when importing queries does not show the correct text, it forgets about all the umlauts – see the screenshots to the right (PMR 61876,212,848).

Two workarounds, neither of them do actually work. No ETA for a fix yet, because it has to be forwarded to Information builders.

Straight from PMR 61859,212,848 – the NLS issue (details here).

ACTION TAKEN: Reviewed log
java.lang.NullPointerException
at _ibmjsp.console._wfconsolenls._jspService(_wfconsolenls.java:91)

Checked open case with IBI, still no fix.

The customer will need to continue to use English until a fix is
available for this issue.

This is not quite what i was hoping for. I had to wait for a full six weeks to get my hands on QU2 after the release dates, while other people in the USA had it a day after. I know that IBM has to forward all this stuff to Information Builders, but it would’ve been really cool if it worked right from the start.

Yesterday i’ve installed 5733-QU2 on a new model 515. While doing so, i ran into the first problem.

Right now, i’ve got two PMRs open, for two different problems:

PMR 61859,212,848

The NLS problem that i’ve talked about during the installation has not gone away. I still can’t use QU2 in German, and the administrative web interface doesn’t work right either. I’ve got a callback from IBM this morning regarding the issue, and was told that the problem was already known, and that they’re working on a PTF.

I know that software testing is an expensive process and also costs a lot of time. But this does not look like an issue that could’ve escaped QA that took the whole internationalization thing seriously. I wish IBM/IBI would’ve spent a bit more time on this, in order to release a product that works completely from the get-go. It’s all about impressions!

PMR 61860,212,848

This is a new problem that i’ve come across yesterday when testing basic functionality. I’ve tried to import a Query/400 query definition, and received an error message when trying to do so:

(FOC1740) EXECUTE ERROR system -q "CALL QZDFMDB2 PARM('-t' '-f'
'/qibm/userdata/WebQuery/ibi/apps/baseapp/BUDGET1_call.txt')"

I’ve even tried to use the query provided with the redbook sample library QWQCENT, but this didn’t work either, same error message. I’ve got a callback from IBM about this issue too, and was told that the problem is currently under investigation, and a fix isn’t immediately in sight, again because IBM can’t actually fix the problem – they have to forward it to Information Builders, which are the guys that actually produced QU2 (as part of their WebFOCUS family).

Other drawbacks

When i heard about 5733-QU2, I thought this to be great news for the System i community. Finally a decent, System i based, graphical query application. But there are some serious drawbacks that you need to consider. I fully hope that IBM will try to improve the situation for their customers, but not all of these points are easily fixable. Of course, QU2 is still miles better than the age-old QU1.

• Separate pricing (named user, instead of the 515/525 concurrent user model)
• Separate software maintenance (Why? I don’t understand this at all.)
• Convoluted installation (requiring about 4-6 hours!)
• The slow speed when working with it’s web interface (i expect sub-second response time on every click for an intranet web application)
• Slow report generation speed
• Usability and looks like it escaped from 1995
• Cumbersome data handling, moving away from native i5/OS integration using *LIBL etc. which was very convenient for *QRYDFN
• Incomplete functionality (e.G. can’t delete unnecessary metadata)
• PDF generation doesn’t work completely (that’s even documented in the redbook)

It could look better, really

If you haven’t worked with QU2 yet, see the screenshot to the right. Does this look like a web application straight from the year 2007? No. It looks like it was made and created in 1995, using CGI programs written in C. The response times are awful – i expect a click to have an immediate impact, with barely no waiting time. This is not the case with 5733-QU2. Especially in the Report Assistant, you might have to wait multiple seconds until a window closes or a new one opens. I don’t find this to be acceptable.

Now, IBM can actually make web interfaces that look kinda decent and 2007ish. As you can see to the right, this is the HMC v7 web interface, which actually looks good. I’ve never used it though.

There’s are several products that come pretty close to my expectations of a decent web interface – Gmail< /a>, Google Reader, and the Microsoft Exchange 2007 Outlook Web Access (nice flash demo), the latter is also pictured to the right. All of these applications are snappy, react in a sub second manner, are fully interactive, have decent graphical dress, support full drag and drop, and integrate nicely into their native environments.

I’m obviously not the only one complaining – New Generation Software, Inc. published a document about the QU2 drawbacks.

Why doesn’t QU2 look like a decent web application from 2007? Why is it so slow? Why doesn’t it integrate nicely into i5/OS’s DB schematics? Why are there such show stopping bugs with basic functionality?

I DO want a decent replacement for Query/400. I DO want our customers to move into the current time. But IBM isn’t making it easy. I hope though we will have fixes for all these open PMRs soon.

Anyone care to share their experiences with QU2? Anyone already working on a roll out plan?