redelijkheid.com

Biggest (techno) news in the Netherlands was probably the failure of the Internet. The largest ATM Internet backbone massively failed in the Netherlands. This resulted in thousands of people without any Internet access for 1-3 days.
Currently the problems are temporarily patched, but they are still on high alert.

For the first time I wasn’t affected. Normally this only occurs to me.

Purely coincidentally, I found a South Park episode (s12e06/Over Logging) in which the Internet fails. This results in mayhem in South Park (just like it did over here).

Furthermore, I drove 170 kilometer today, and it took me about 5 hours. Traffic jams due to spring rain (????) were to blame (according to the traffic information services). I think it’s because people are egoistic when they’re driving, and won’t anticipate (when they’re reading the paper, shaving, and/or applying make-up using the rear view mirror).

So no Internet (for large parts of the Netherlands), and traffic jams from here to Tokyo. So slow might be an understatement.

Ever since I upgraded to Wordpress 2.5 I had to moderate every comment posted on this blog. Note that I hadn’t turned this on in the Wordpress settings.

Some research showed that it might have to do something with ‘older’ plugins. After a trail and error period I found out that the WP Hashcash plugin was to blame. Ever since I de-activated it, the problem seems to be gone.

Well, that was another morning well spent….

A couple off weeks ago I started experimenting with FreeRADIUS on Ubuntu server (v6.06.2 TLS). Mainly because I needed to test some things for work. So I used VMWare to experiment. The networking part was set-up as Bridged.

Today, I wanted to test with iperf (a tool for network performance testing). So I launched the virtual machine, but there was no network connectivity. ifconfig showed that eth1 didn’t received an IP adres.
So I ran every possible test there was;

• restarted the interfaces (/etc/init.d/networking restart)

This resulted in the following;

Listening on LPF/eth1/00:0c:29:68:e3:eb
Sending on LPF/eth1/00:0c:29:68:e3:eb
Sending on Socket/fallback
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 8
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 12

The “DHCPDISCOVER” messages continue about 4 times, then the message:

No DHCPOFFERS received.
No working leases in persistent database - sleeping.

• restarted the DHCP server
• rebooted the virtual machine
• changes the networking to NAT instead of Bridged (this way, connectivity was restored, but not the way I wanted. I needed Bridge-mode)
• Tried to run the virtual machine on OSX (VMWare Fusion), which worked surprisingly.

After this I ran Wireshark on my server to see if DHCP request were coming in…. And you might have guessed; No DHCP request were reaching the DHCP server. So the problem was work PC related…. As a matter of fact, I had the Cisco VPN client running….. Which didn’t allow the DHCP request broadcast.

Shutting the VPN client down solved the DHCP problem. After the virtual machine worked I could reinitiate the VPN.

Mental note to myself: do NOT boot/restart the virtual machines when the VPN is up.

Since I upgraded from OSX 10.5.1 to 10.5.2 I haven’t been able to use my shared printer on on my Windows server. Every job I submit gives me an error saying On Hold (Authentication Required).

There’s no way that I could submit the printjob to the server from my Mac.

It seems that there is a bug in the authentication mechanism on OSX 10.5.2. If you add a (network) printer by using the System Preferences, you get to browse to find the actual printer. this results in a printer in the following location:

smb://<servernetbiosname>/printershare

For some weird reason this won’t work. You need to specify the IP adres of the server or include the windows workgroup name in the url. This can be done by using the System Preferences. But you can also use CUPS. This is the webinterface on OSX to access printing services. You can access CUPS by entering the following URL in your browser;

http://localhost:631

This give you a webinterface for managing printers. Just add (or edit) the printer. During the ‘Wizard’ you can specify the actual path/URL where CUPS can find the printer. Just one of the following:

smb://<username>:<password>@<IP Address>/<Printershare>

smb://<username>:<password>@<workgroup>/<servername>/<Printershare>

Don’t worry about the username and password being displayed in the URL. These will be ‘moved’ upon insertion to the Keychain.

Alternative is to add the Advanced button on the Add Printer Wizard on the Print & Fax panel in the System Preferences. This can be done by doing the following:

• Open System Preferences
• Open Print & Fax
• Click on the ‘+’ to add a printer.
• Right-click (or ctrl + left-click) on the menubar
• Customize Toolbar
• Add the ‘Advanced’ button

With the Advanced button you can specify the exact smb path mentioned earlier to add the printer by its IP address or with the workgroup information in its path.

More info on the Apple Support Discussion pages [1, 2].

Early this week, I found a new firmware for my Nokia E61i (out since October 2007). The version I had was v1.x, and this one was v2.0633.65.01 (press *#000# on your E61 to see the current firmaware version).

Updating goes through a separate application, but it should also be possible through download over the mobile network (I haven’t tried this). All you need to know is explained on the Nokia website, but there are some thing you don’t want to forget;

• Use the Nokia Datasuite to create a FULL backup of the phone, because during the upgrade the phone goes back to factory defaults.
• Have lot’s of patience (and some deodorant handy)
• Make sure the PC isn’t doing anything else that might interfere with the update.

The first attempt failed for me. Even with all the warnings (DO NOT INTERRUPT THE UPDATING PROCESS OR DISCONNECT THE PHONE!!!) I rebooted the PC and disconnected the phone. Result, the phone didn’t respond (this is where my deodorant came in). After this I restarted the upgrade process, and the phone got recognized (thankfully). After 10 minutes, the phone rebooted with the new firmware version.

First action was to restore all data and settings on the phone. Since I had some issues with my network connectivity I decided to remove the Access Points and reconfigure them…. Well, don’t. First of all, I wasn’t able to receive the configuration parameters from the mobile operator (SMS ‘ja’ to 1300), so I had to reconfigure them manually.
Their website has a step-by-step manual on configuring the E61i, but this didn’t help either.

MMS functionality remained absent, and none of the applications was able to connect to the Internet by itself. I had to initiate the connection manually before browsing the web.
E-mail was even worse. Every time I had ‘Packet Authentication’ errors. So after a day I decided to restore the Access Points from my backup, and guess what… Everything worked again.

And now for the thing that have changed (at least the ones that I’ve noticed);

• the e-mail application seems more stable
• camera and video are still crappy (the time between the snapshot sound and the actual capture is still multiple seconds)
• Idle connections are terminated. This is a bad thing (for me at least), because I had my UMTS connection open all the time, and this way I received mail throughout the day. Now I have to connect each time I want to check my e-mail.

I haven’t tried VoIP yet, but will try to do so soon (that’s what happens when you don’t pay for your own bills )

Conclusion is that the phone didn’t get better. There are some improvements, and there are some new annoyances. But my e-mail is stable at this moment.

Since I have an iMac with OSX 10.5 (Leopard), I use TimeMachine for my backups. This works great actually. But I also need an off-site backup of some sort. Just in case the house burns down or that some f*cker decides to steal my hardware.
So I bought an external Freecom 160GB USB2 drive (USB powered) for my off-site backups.

I encrypted the entire harddisk with TrueCrypt 5.0 on my iMac, and copied the data I needed to preserve. After that I wanted to access the data from my work laptop (Windows XP SP2 with TrueCrypt v5.0)….. This didn’t work. TrueCrypt didn’t recognize the password, or the encrypted disk (AES / SHA-256 full disk encryption).
I tried to access the data on my Mac and everything worked, so there’s no data corruption of some sort. Eventually, I recreated the encrypted drive on my Windows XP laptop (lost the backup in the process). This time the disk would mount, and could also be read/mounted by my Mac.

So, I guess that TrueCrypt is Cross-platform, but with the current version (v5.0a) you need to make sure to create the volume on Windows if you also want to mount it on OSX.

I reported this through their bug-reporting tool to the developers. No idea if there are similar problems with Linux.

UPDATE: Pretty soon they released v5.0a, and today v5.1 was released. So development goes on

As you might have noticed, the server was down the last 24 hours. When I got home last night I found my server dead. The green power LED on the mainboard was still on, but apart from that… nothing. So I tried to switch the server on and off a couple of times. After a few tries, the entire room lit up for about a nanosecond followed by absolute silence, and darkness…..

The power supply had short-circuited and took the entire house with it. Byebye power supply.
The initial thought was; ‘Well, it’s gonna be a nice weekend reinstalling the server from scratch’, but thankfully, it was only the power supply that had died on me.

After replacing it, the server ran as before. This got me thinking about the life-expectancy of computer hardware. Frequent visitors may have read a similar post on my blog last year. So the power supply that died isn’t even a year old. I still might have had some warranty on it (if I hadn’t pried it open to see the internal damage ). If I would like to use the warranty, I’m looking at at least a couple of days downtime, and since a new power supply is relatively inexpensive, I couldn’t be bothered.

The life expectancy of hardware is obviously shorter than the Mean Time Between Failure (MTBF) given by the manufacturer. The new supply has a (theoretical) MTBF of >100.000 hours, This means >4000 days. Well, mine lasted <365 days.

I bought some accessories for my photo gear today. Among the items was a new 2GB SanDisk Extreme III CF card. This package came with a ‘purse’ to store cards, and a CD with recovery software. In most cases you’ll only get the Windows version of a product (if there is an OSX version all together)….. But this CD also contains the OSX version of SanDisk’s RescuePro. Downside is that the CD is mini CD (or what you might call it), and those won’t work very well with the latest Mac’s. The latest Mac’s have DVD/CD players without a tray, so insert the CD, and you can ship your Mac for repairs……

And since the Windows OS can’t read the Mac part of the CD, there’s no way of getting the software of the CD.

One thing I hate is not knowing where I am. I had this in Windows, and I still have it OSX.

By default OSX doesn’t show the full path in the finder. This is annoying as hell. Especialy when you’re having shares called ‘Pictures’, and the chances are quite big that you might have a local directory called ‘Pictures’. You can’t tell the difference.

Using the following terminal command, you get the full path in the Finder. So no more ‘Pictures’, but ‘/Volumes/Pictures/’

Terminal command:

defaults write com.apple.finder _FXShowPosixPathInTitle -bool YES
killall Finder

Simply replace the YES for NO to disable the feature. This is OSX 10.5 or higher only!!

I evaluated the MDaemon v9.7.2 mailserver, and I must say that it’s quite complex to get it to run… my way.

It’s quite common to separate the program and the data. So, I tried to put the data on a different disk. This isn’t done by using the GUI, but you need to edit the Mdaemon.ini file in the App directory (not every directory can be changed by the GUI). After changing the defaults, the MultiPOP feature didn’t work. It seems there’s another config file for remote mail stuff. This file (schedule.dat) needs to reference the same location as the Mdaemon.ini file. Why is that? Why use two config files for one program? And why use a different extension (.dat versus .ini)? Beats me.

For a mailserver that pretends to be a fair, and relatively simple alternative to the over-powered and over-featured Microsoft Exchange mailserver you need a ‘rocket science engineers degree‘ to get it to work properly.

I’ve been looking for a GPS logger for a couple of months. A GPS logger can be used for attaching GPS information to your digital photos. This way, you can display your photos on publicly accessible maps in Flickr or Google Earth.

More and more of those GPS loggers are surfacing on the market [Sony GPS-CS1 / GPS-CS1KA, Globalsat DG-100, Qstarz BT Q-1000, GiSTEQ PhotoTrackr, WBT-201]. But they all have one problem; None of them seem to work (natively) with an Intel Mac running Leopard (source). Some of them require open-source drivers to function. The downside of that is that an update from Apple or an updated firmware in the GPS logger might cripple the functionality.

Why is it that GPS manufacturers won’t create a GPS (logger) device that simply works on both platforms. It couldn’t be that hard. Just create a GPS device that is also recognized as a external drive, containing the raw GPS data in the most common format (e.g. NMEA). This way even the Linux users are not left out.

Having a day off from work meant that I had some time to clean-up my server from everything crap. This also meant trying to upgrade Coldfusion MX 7 to Coldfusion 8. The new version supports some nice new features, with which I wanted to experiment.

Since none of my production websites run on Coldfusion (currently PHP), I could safely uninstall and install the new developer version. When I pressed ‘remove’ I had this flash from the past. When I tried to install version 7 I had humongous problems with the installer (services didn’t register correctly with IIS, etc.). But this could be just a coincidence…

Installing Coldfusion itself was straight-forward. The fun began when the installer wanted to launch the admin-panel to finalize the installation. No responds from the webserver. Even my blog was not working (”System cannot find the file specified“). Seemed that there were references to the old Coldfusion parser in the IIS.. After removing those instances, I had my blog back up-and-running.

Next was to finalize the Coldfusion 8 installation. I had to manually add the Coldfusion parameters to IIS (allowed Web Services, document types, … the works) After screwing around for nearly two hours, I could finish the installation.

Just to make sure that it wasn’t me, I installed it also in a virtual machine. And off course it worked flawlessly. Guess I need to reinstall my server in the next couple of weeks. Who knows what else is f*cking up the system.

Oh, another thing; I used to use RDS for development of the websites, so I tried to configure it again…. Well you might guess it… Not working. After yet another hour searching I found that RDS uses a ‘virtual file‘ called ‘ide.cfm‘. This file is located in /CFIDE/main/. The actual location (main/) doesn’t exist, so IIS returns a 404 error.
It seems that you need to uncheck ‘Verify if file exists‘ in IIS for the .cfm extension for RDS to work. Or do as I did; create the directory ‘main‘ and create an empty file called ide.cfm in that directory.

Well, this was an afternoon well spent…. NOT.

Next mission is to upgrade MySQL to version 5. So expect some serious downtime.

UPDATE: MySQL has also been upgraded to v5.something. It went surprisingly easy.

Well, nothing is as easy as it may seem. As I mentioned earlier, I downloaded a copy of Leopard to see if it’s worth buying. This download was a so-called .dmg file. This file can be burned to a double layer DVD by the Disk Utility or by using Toast.

The problem is that my MacBook Pro doesn’t have a dual layer DVD burner, so I need to burn this one with my external USB NEC burner (which supports burning dual layer DVD’s). For some weird reason it (my MacBook Pro) refuses to burn anything dual layer. So in the mean time I’m searching for another version of Leopard (preferably an ISO-file), and I’m using Super Duper! to backup my MBP. Both are taking ’some’ time, so I don’t think that I’ll be upgrading this weekend .

UPDATE: Oke, I found a PDF from Adobe stating that Adobe Lightroom isn’t fully compatible yet. So, no upgrade for me. Guess I can stop trying to get the downloaded DMG to work. Well, at least I didn’t spend my Sunday on nothing. I figured out how to convert a DMG to ISO.

Most of the TV series I watch, I watch on my MacBook Pro. I just hook up the laptop to my 40inch HDTV, and enjoy Stargate Atlantis, Heroes, Lost, House, etc. in HD quality (mostly 720p, because the MBP isn’t fast enough to decode 1080p ). But the only way of enjoying this to the fullest is with Dolby Digital or DTS, but Quicktime doesn’t support this. Only Dolby Surround is supported by Quicktime.

Last week I installed Perian. This is a so-called codec pack which enables Quicktime to play more video formats. After installing Perian, it’s even possible to play Matroska files from within Frontrow, but still no DD or DTS support…..

I wonder why Apple hasn’t added support for DD or DTS. It’s shouldn’t be too hard (the DVD player does support DD and DTS). Especially, since Apple is the choice for content-creators…. I guess VLC is here to stay for a bit longer.

I went to Tibet in 2005. This was one of the best holidays ever. Only Africa could beat it in my top 10. The country is just beautiful. It’s just a damn shame that the Chinese are ‘corrupting’ the country with modern transportation, and genuine Miss Tibet contests…… What’s so Tibetan about this???