Henri Cook

I’ve been messing aroung with songoftheminute.com since it launched a couple of days ago – i’ve slowly found myself using it more and more for ‘general listening’. As the name implies a different song is oftered up to visitors every minute but for me the added value comes with how those songs are selected – the site uses last.fm’s api to find its songs but it only selects songs which it’s algorithm deems to be ‘popular’, the result of course is that I get the best music in my chosen genre in a great random playlist. Easy to use anywhere and handy bio information for when you’re just surfing, two tumbs up

Nokia Nokia Nokia, i’ve been a faithful customer for years but the Ovi store might be the last straw.

I’ve had problems with you spamming my email in an impossible-to-unsubscribe manner, your poor online customer service and you’ve slowly been falling behind as the IPhone has become more and more prominent.

I’ve just logged in to the Ovi store all excited and i’ve come out with one application – SmsPreview, which looks and works great. The website crashed a few times (time outs) before I got it to load, the buttons have no user feedback and aren’t normal html submit buttons so I had no idea when I was actually submitting stuff which lead to a few double-submits and some confusion during signup.

The ‘20,000 apps’ that you’ve launched with are mostly bollocks themes and ring tones (hence me only finding one that’s of any use to me). I can’t believe you’re counting a star trek trailer as an ‘app’ – i’m really disappointed at the trickery. The internet market isn’t stupid – in fact it’s probably the hardest to fool of them all, this is a really bad start for Ovi.

All you needed was a private beta, ask 200 users like me to login and use your store for 30 days. I could have told you the forms were rubbish, it was unclear how to select what phone you have and that having themes as apps was a crap idea. I could also have told you that you need to give users the ability to comment on the apps they download as well as rate them.

I hope this gets better, next month’s going to be a big one for phones and I for one am looking for a new one. The N97 looks nice but I need Ovi to be good and I need Symbian to be a responsive, intuitive, *well tested* operating system unlike it is at the moment with my current N95.

I was surprised this morning to find that McAfee ‘On Access Scan’ had deleted my copy of Spotify overnight saying that it was an example of W32 IRCbot.gen.z – thanks Mcafee, you’re now disabled.

This of course raises an interesting question – does Spotify connect to IRC somewhere? What other reason could McAfee have for detecting it as a chat bot in this way?

After a morning looking at phones to buy the ideal combination to me would seem to be the Nokia N97’s hardware running Android – alas such a combination doesn’t exist, if I want Android I have to put up with the ugly HTC G-Phone range. Iphones are getting a little tired and I miss things like contact groupings and MMS so they’re a poor option (when comparing to my existing N95). The Blackberry storm looks nice, but I really like Android as an ‘open’ operating system – there’s also the pain of getting the bloody things unlocked.

This lead to some reflection on the current length of mobile contracts e.g. 18 months or 24 months (Orange.co.uk) – for those of us who remember it used to be 12 months, and that was considered a long time. I would think that, because of the extended minimum length of contracts from the Mobile Phone operators, phone providers are going to be forced to push down their prices for sim-free handsets. No matter how hard I try I doubt I could make a new phone last two years, i’m going to need to buy a new one before then and I can’t ‘abuse’ my operator’s free upgrade any longer. Yes, this does mean operators can expect more ‘direct’ business (i.e. not through mobile operators) but at the same time they’re going to be experiencing a loss as the operators buy less from them. People like my mother for example would get a free upgrade simply to sell it on on a website like ebay,  not actually needing a new phone – these people are going to stop doing that with longer contracts and companies like Nokia, Samsung, Blackberry and HTC are going to suffer. Hopefully this will lead to phones that aren’t more expensive than a high-end PC.

I was recently asked to repair a five year old Sony Vaio PCG-FX705, it’s got a 600 Mhz processor don’t ya know? Anyway it overheats and shuts down a lot which makes it bin worthy to my eyes, but a sentimental owner has a different opinion. Here’s a couple of pictures of my solution to the overheating problem (while I got Windows to install) – many thanks to Kopparberg and Carlsberg for their assistance..

I’ve just upgraded from Intrepid to Jaunty. In the past updates have often prompted a system reinstall to get my relatively unsupported nvidia graphics and Creative X-Fi working again.

At least this time no reinstall was necessary, my install is still plagued by several problems that Jaunty didn’t spot automatically (and in fact reset from my intrepid settings):

Nvidia Chipset On-Board Gigabit ethernet (MCP55)

I still have to add the following to /etc/modprobe.d/options.conf

options forcedeth msi=0 msix=0

Without this I had no internet connection after the ‘big reboot’. I would assume this is the forcedeth driver guys fault, so shame on you.

Nvidia Graphics Drivers – Unbroken, well done!

X-Fi drivers still unsupported

I had to recompile the XFi GPL2 driver and reinstall it for the new kernel.

Encrypted Root Partition fails

Although my encrypted root partition appeared to mount it failed when trying to mount what was presumably swap. I’m still exploring this problem but found one can skip past the busybox error (ALERT: could not find /dev/disk/by-uuid/XXXXX) by using ctrl+alt+sysrq+letters in order(RSE).

Overall not bad, it’s a really nice looking release!

I recently co-located a new server to find that I was given one ‘main ip’ on subnet X and all my additional IPs were on subnet Y. This has happened in the past but the Datacentre setup usually has a gateway on each subnet to allow for internet access. This was not the case in this instance, instead the additional IPs were expected to use the host machine as a gateway (bound to lo:* addresses). In my previous KVM setups i’ve had br0 as a bridge over eth0 and all guests have received unadulterated access to the network to use the standard network config (i.e. the same gateway as the host machine).

In this case that didn’t work (as machines with their primary IPs on subnet Y could not access the subnet X gateway, and there wasn’t one for subnet Y).

Instead with a lot of help from the Bluesquare support team we were able to come up with a solution, using a virtual device (br:0) bound to one of the IPs in subnet Y to act as a gateway, effectively bridging the connections between addresses on subnet Y with the main address on subnet X. This is best explained by showing you my /etc/network/interfaces on both host and example client, both running Ubuntu Linux (Server ed.):

Throughout please assume 192.168.1.100 is the machine’s main IP (on subnet X) and 192.168.2.xxx is subnet Y with 192.168.2.101 being the additional IP chosen to act as our ‘virtual gateway’:

On the host:

# The primary network interface
auto eth0
iface eth0 inet manual

# Standard bridge interface on the main IP
auto br0
iface br0 inet static
address 192.168.1.100
network 192.168.1.0
netmask 255.255.255.0
broadcast 192.168.1.255
gateway 192.168.1.1
bridge_ports eth0
bridge_fd 0
bridge_hello 2
bridge_maxage 12
bridge_stp off
# Virtual Bridge for additional IPs
auto br0:1
iface br0:1 inet static
address 192.168.2.101
netmask 255.255.255.0

On the guest:

# The primary network interface
auto eth0
iface eth0 inet static
address 192.168.2.105
netmask 255.255.255.0
network 192.168.2.0
broadcast 192.168.2.255
gateway 192.168.2.101

auto br0:1
iface br0:1 inet static
address 192.168.2.101

In the Libvirt Configuration file (or convert to command line if you so wish):

<interface type=’bridge’>
<mac address=’00:16:3e:01:00:15′/>
<model type=’e1000′/>
<source bridge=’br0′/>
</interface>

Finally:

Make sure “echo 1 > /proc/sys/net/ipv4/ip_forward” has been executed – IP forwarding will be needed for this.

Conclusions:

I actually like this approach, although it wastes one of my paid for IPs it allows all my guests to use one of my own IPs as a gateway. If for any reason I ever need to change the gateway of the host, I no longer have to go through each individual client and change the settings there as well.

Please post any comments/questions.

Our university firewall is shockingly restrictive, this meant that recently I couldn’t even download a game that I bought using Steam (this is my first time using the Steam system).

There instructions will also work for any program you might be unable to use behind a firewall in a university that’s still old-fashioned enough to limit its residential users in this way.

1. Install Tor (https://www.torproject.org/download.html.en) – Standard configuration is fine

2. Tor is now listening on 9050 for SOCKSv4 connections, install SocksCap (http://www.brothersoft.com/sockscap-192885.html) – Standard Configuration, setup the SocksV4 proxy on port 9050 when prompted to provide Sockscap settings.

3. Add the ‘Steam’ launcher to Sockscap

4. Run Steam, optionally keep the Tor window with the bandwidth graph open to confirm everything is working.

This is a very simple guide, I haven’t gone into the ’step by step’ configuration of any of these components – if you can’t figure it out you probably shouldn’t be doing something so complicated.

Typically you can get ~40k/s over the Tor network, i’m looking into the ability to provide one’s own Tor proxy (with privileged bandwidth for the owner) – more soon!

Nevertheless if you have any questions. please post them here.

As of the time this is written this is currently impossible – sorry! If anyone finds out how to do it please let me know!

After months of aggravation and ‘just putting up’ with this bug a (source) patch and a quick terminal-based fix have been released.

Bug comments/description here: https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-177/+bug/296167

The terminal-fix which I’m using until a patched package is released is instituted as a script, and can be bound to a gnome hotkey for super quick-fixing.

/home/henri/scripts/mouse_fix.sh:

#!/bin/bash
cat /home/henri/scripts/mouse_fix.sh
xte “mousemove 0 0″

For this command to work you need the xautomation package installed.

Many thanks to ‘geekfreak’ from the link above for coming up with this fix!