Going Solar

Flattr this!

I’ve been interested in going completely solar for a while now back home in South Africa, as pricing for electricity has rapidly increased past the pricing for solar; return on investment is in the 3 year range currently.
It will get close to 1 1/2 year return on investment when Eskom new pricing increases happen, so its a no brainer to install.

I’ve already replaced our geyser (hot water system) with a solar based system, plus all the lighting in the house is already LED based (yay China!), so our base load of electricity is low for the size of the house. I can still improve though by installing solar, to make the electrical costs approach zero, and at some indeterminate point in the future when Eskom allows for legalized grid tie, a profit center!

As I’ll nominally be a 10KW producer (I can add 2 panels to get there), I should be able to at some point pump back into the grid sooner rather than later – as the trial projects for Cape Town all sit at the 10KW range and up…

From what I read –

Eskom will pay out R 1.20 per kilowatt hour
generated by your solar system for the first three years, 70% immediately after installation
and the balance at 10% for the next three years.

More here –
(Yes, currently this is only available for commercial use, but I do expect that to change at some point. My system should be in use for at least 25 years, so I should get some benefit at some point in the future.)


Before I get to pricing though, I need to explain how it all works.

For any system, you’ll need some kind of input.
As I’m looking at Solar, thats my input. I can also look at thermal or wind based. Wind based is a distinct possibility in Cape Town, but I have been advised that its probably too windy to use! (turbines can’t run during extremely windy weather or you break the turbines).

So, I’m going with Solar.
There are 2 types of solar panel out there. Monocrystalline and Polycrystalline.
Monocrystalline is more expensive per watt, as its a more difficult process to make panels from.

Mono panels are also slightly smaller per watt of output. On average mono panels are about 14% smaller. They also work better in hot climates.
Aside from those differences, they’re fairly similar.

Panels are typically rated in watt terms.
A 300w panel will give you 300W of power at peak output (eg mid-day).

This 300w of power is at DC voltage though, and for house use, we need A/C

The 300w panels I’ve been looking at give 36V @ 8.3A.
I’ll probably go with polycrystalline, as the pricing isn’t really worth the extra 30% for mono crystalline for my needs.

Panel info below – (click me for pdf)

Screen Shot 2013-05-13 at 5.55.14 PM

Basic calculation for power output is P = V * A

This works out to 300W a panel (303W = 37.6V * 8.06A)
I’ll be getting 30 panels, as thats about the max I can fit on my roof in theory.

Screen Shot 2013-05-13 at 5.19.09 PM

(My brother hasn’t gotten me the exact sizing yet).

To use this, we need an inverter though, as something has to convert the DC power into AC.

In my house, I have 3 phase power, and an antique metering system.

Screen Shot 2013-05-13 at 5.08.01 PM

3 Phase is good, as i have sufficient power for my needs, but its bad as I need a more advanced inverter to give me 3 phase.

I could use 3 x single phase inverters, but for simplicity, I’ll be going with a single 3 phase inverter.
If you see the cabling here – you’ll see we have 3 phases + 1 neutral = 4 cables.

Screen Shot 2013-05-13 at 5.09.15 PM

To work out what sort of inverter I can use, I need to do some basic math.

I’ll have 30 Panels total.
Each panel gives out 36V @ 8A., and that will give me approximately 9KW output. As the smallest *decent* 3 phase inverters I could find are 10KW ones, thats a good size.

I have a choice of running the panels in series or in parallel.
If I run them in series, then the Voltage increases.
Eg 1 panel = 37v, 2 panels = 74v @ 8A…

If I run them in parallel, then the Ampage increases.
Eg 1 panel = 8A, 2 panels = 16A @ 37v

If you’ve ever seen welding cables or car battery cables, you’ll see what sort of cabling is required for high Amps. So, everyone wires using DC voltage.

My inverter of choice is probably going to be this: Growatt10000UE

Screen Shot 2013-05-13 at 5.12.58 PM

That 3 phase inverter has the following characteristics.
It will power up from 300V (min voltage to run), and accepts voltage up to 1000V.
It also has 4 inputs for panels.

Generally each input is called a “string”.

As I’ll have 30 panels, I’ll probably be balancing them out in 2 x 15 piece strings -> the inverter.

Each string will work like this

37V * 15 = 564V DC * 8A (4.54KW of power)
37V * 15 = 564V DC * 8A (4.54KW of power)

This will give me a rough total of 9W peak power.

As conversions are never perfect, and panels can output more during peak than they are rated for, I’m getting a 10KW inverter. This will allow for some small headroom in future if I need to expand slightly.

It also is fine for something I haven’t talked about yet – open circuit. The panels I’m looking at run at 36v open circuit (i.e. before they kick in), the inverter also needs to be able to work without issue at open circuit voltages. As the inverter supports 1000v, open circuit of 564v isn’t an issue.

So far, costs are:

30 Panels = 720RMB / poly panel = 21,600 (mono panels are about 900-1000 per piece). Poly panels are physically 1.9M x 1M @ 300w / 28KG , Mono 1.9M x 1M @ 300W / 25kg
10KW 3 Phase inverter = 9,000RMB
Weight = 1000KG with packing.
Shipping + clearance – roughly 15,000 + duties @ 20%

Total landed in Cape Town = 45,000RMB / R60,000

That gives me a rough pricing of R6.6 a watt *installed*.
It also gives me a system that I can hook into the grid (illegally currently!), but won’t provide for power in case of failure.
I actually don’t need something that size, but sadly, due to the cost of clearance being a complete rip off, it doesn’t make sense to ship less 🙁

Currently our power bill sits at about 700-1000 rand a month, over a year this is around R12,000 using worst case scenario maths 😉
My intended system will cost me about R60,000 + install labour. At current electricity pricing, I should see a complete payback for the system in about 5 years. Given that electricity prices are going to be *doubling* over the next 5 years in Cape Town, this will actually be achieved in about 3 years or so.

Not too shabby!

Our current monthly electricity usage looks like this for those who may be interested.

Screen Shot 2013-05-13 at 5.25.54 PM

You’ll note that electricity use spikes on certain days (mainly weekends) – this generally ties into when the maid is there, as then the washing machine, dish washer etc get run, or on the rare occasion that my brother actually cooks 😉

Initially I’ll be feeding excess power back into the grid, and using that as a “battery”.

How will that work?

Well, as I have an older meter, it can run backwards. So, daytime when I have _substantial_ excess, i’ll be running backwards, and nighttime, when the solar panels are not generating, I’ll be running forwards.
Essentially, using the grid as my battery..

Eskom will be benefiting from all this, as I’ll be a net producer far over what I consume – so they’ll get all the free electricity I’ll be generating.
Its also safe – as the inverter will not feed back into the grid if its offline – eg when we have one of our rather too regular power outages (3 in the last month from my logs).

Longer term I’ll be installing a battery system to allow for complete off grid, but funds don’t currently stretch to that yet..

Do note that the above is for my needs – your needs might not be my needs!
I need a 3 phase system. Most people _don’t_. I’m also going grid tied for the moment due to funding available. Others might find it better to have a hybrid grid tie/ battery system. If I could afford it, I’d go that route!

I’m also *heavily* overspeccing the output – clearance costs are substantial for South Africa (highest in the world almost), so it doesn’t make sense for me to ship a small system, as there is only a marginal cost for what I’m speccing.

A suitably sized system for us would be 8 panels, and a 3kw inverter. I’d be crazy to ship that though, as the clearance is more than the cost of the system. So, I’m heavily overspeccing on requirements so that it makes sense. Long term its also a no-brainer for me, as I’ll have substantial excess I can sell back to the grid.

In case anyone is interested how I’ll retrofit this sort of system with a battery backup – here is a diagram of a single phase implementation – I’d be doing something similar:


That said, I do have another easier solution – I’ll probably go cheap – stick the things that may not lose power(tm) circuit on a 2KV UPS, and have an isolator switch in circuit for when the grid goes down so its isolated from Eskom. This will accomplish the same thing pretty much, and should tide us over for the average 3-4hour outages we seem to experience every few weeks. It will also sit nicely in the computer rack that will contain the media side of the house and data storage needs 🙂

Flattr this!

Debian 7 was released, and like all good sysadmin’s we updated our servers to the newer OS version.
Unfortunately there have been a few minor issues where functionality has changed.

One of those relates to FTP.

We generally use VSFTP as its relatively secure, and has chroot functionality out of the box (i.e. users can only see their own folders).
Today a user complained that they couldn’t login, and on checking we saw this little error:

500 OOPS: vsftpd: refusing to run with writable root inside chroot ()

A quick google showed that 2.3.5 breaks backward compatibility somewhat by changing functionality, oops!
This was done for safety reasons, but is a bit of a questionable change to functionality when it breaks things.

The suggestions on the internet were to add this line:


Which didn’t work – VSFTP wouldn’t restart, as it didn’t recognize it. What to do?

Well, the quick solution is to use Debian unstable, as that uses a newer version of VSFTP 3.x which does recognize that line.


(assuming you have the testing repo’s in apt.sources)

apt-get install -t testing vsftpd
echo allow_writeable_chroot=YES >> /etc/vsftpd.conf
/etc/init.d/vsftpd restart

Flattr this!

As one of the main contention points people have with mail service is either the amount of spam they receive, or the amount of legitimate email we block, we’ve decided to put the solution in your hands.

We’ve added user access to the blocking implementation we use at Computer Solutions.

For a quite rerun on this our incoming mail rules are as follows:

  • Sending Server has a valid Reverse DNS Entry
  • Sending Server conforms to mail RFC’s
  • Sending Server is not listed in any of the following Antispam Service Lists
  • Mail does not contain a virus, malware or similar content.
  • Mail is addressed to a valid sender.
  • Recipients mailbox is not full.

We’re giving you access to do what you want with regards to incoming spam blocks.
If you decide that our heinous blocking of senders who’s servers are _definitely_ listed in spam listings is not to your taste, then you can change that.

If you want to whitelist any incoming mail you can do the following:

1) Login as the postmaster account for your domain at (in the example below, I’m editing my own account, you’ll need to use YOUR / password!)

Screen Shot 2013-04-24 at 8.31.18 PM

2) Select Domain Wide Focus

Screen Shot 2013-04-24 at 8.31.47 PM

3) Click Add a domain specific rule (this will apply to all messages received for your domain – i.e. anything

Screen Shot 2013-04-24 at 8.31.56 PM

4) Setup appropriate rules (there are a number of options – in the example below I’m whitelisting all incoming mail).

Screen Shot 2013-04-24 at 8.32.13 PM

5) Note that the System rules below are now greyed out (assuming you whitelisted as per example above).
Thats because they no longer apply!

Screen Shot 2013-04-24 at 8.35.39 PM

In future we will be pushing clients to use this interface for their unblocking / blocking requirements, so that the needs of the few outvote the needs of the many, and your incoming email can go where no wo/man has gone before.


Flattr this!

Over the last few weeks, we’ve been noticing an increase on hack attempts on wordpress installs and other CMS’s (eg joomla).

Most of these attack attempts are from Russian IP space (typically Ukraine), although there are also a lot of botnet attacks from hosed windows computers also (these come from a variety of countries).

To counter this, we have been pro-actively implementing a number of different mitigation solutions, ranging from upgrading clients CMS installs and adding captcha plugins where possible to prevent brute force password attacks, through to scanning for vulnerable files throughout all clients website, and updating them to non-vulnerable versions (timthumb.php being the major issue/problem child that we’ve found to be vulnerable/exploitable).

We have also implemented server-wide lockout systems for failed logins for wordpress using one of our existing protection mechanisms (fail2ban).

Some of you may already have noticed an additional question or captcha being asked during login to your systems.

(example below)

Screen Shot 2013-04-12 at 10.42.02 PM

This is for your safety – if someone hacks into an install, they typically then attempt to run additional items within an install such as malware.

We also have live monitoring for malware running on all servers, and have been quite proactive in upgrading installs which are capable of being compromised.

In the case of a site being compromised and malware being dropped into the site, our live scanner sends us an automated email and we actively investigate.

If we cannot resolve the immediate issue, and find the security hole, we disable the clients site and inform them of an issue, and the need to take further action.
(To date, we haven’t had to go that far though).

We’re not the only people seeing this, although its not well known outside of the web hosting community at this present time.

We believe in proactive solutions for these kinds of attacks, and our multilayered approach appears to have spared us from most of the problems facing others at this time.



Flattr this!

We often have clients that complain about wifi in different area’s of their apartment being weak.

While there are a number of solutions to that, such as running additional cabling, adding additional routers via WDS, using 电力猫 (ethernet over power) adaptors etc, we decided to try out a newish all in one solution from TP Link.

TP Link HomeAV packaging

This ties together 2 paired mini wifi routers via power over ethernet (HomeAV).
See the Chinese diagram below for a clear example.


What does it do?

Well, you can plug one router in where your internet connection is, then plug the other router where your signal isn’t 😉 With a bit of luck, it will connect successfully over your power lines automagically, and there is no configuration needed other than adding in your internet settings at the router side.

The units provide both Ethernet and Wifi at both ends, so they’re a good solution for both wired and wireless access.

The units we bought only have Chinese firmware at the moment, but its pretty much set, and forget for most users.

They also come reasonably well documented – each unit has a sticker with its ip address and user/pass. Yay!

TP Link HomeAV

As you can see below, one unit has 2 ethernet connectors, and the other has 1 ethernet connector.

TP Link HomeAV

The unit with 2 connectors sits at your modem side (it can also extend an existing network if you have another wifi router already).
The WAN port plugs into your modem or network (as appropriate for your needs).
The LAN port provides a network port for use.

The other unit plugs in elsewhere (eg in a suitable location where you need wifi or wired connection).

I initially thought our initial unit was faulty, as it would only pair to the second adaptor for a few seconds, then stop pinging its ip address ( as per its label)

TPLink HomeAV adaptor

Turns out that once its booted up and paired, it gets an address via DHCP from the main
device, so whats written on the back of the unit no longer applies. Stumped me for a few seconds till I realised and went doh!

Here is our suitable test bed.
TPLink HomeAV adaptors

I’m pinging one device from the other –

..and it works.

The units are a bit flimsy, but seem to work fine in our limited testing.

Price – RMB258 + shipping for a paired set via online store (price correct at time of writing)

Extra units – RMB133 (you can have up to 7 units tied together to extend your network).

I’d give these a rating of 7/10 – they’re cheap, and they work.

Further reading (and tools):

Flattr this!

As most of the posts out there are horribly outdated, or provide incorrect information for the current versions in use, here are my quick notes on setting up a time machine share.

First up –

apt-get install netatalk

Check /etc/netatalk/afpd.conf has something similarish to this:

# This line goes in /etc/netatalk/afpd.conf
- -tcp -noddp -uamlist,, -nosavepassword

Add a line for your required shares into /etc/netatalk/AppleVolumes.default

# Time machine share
/nas/backup/timemachine "TimeMachine" cnidscheme:dbd options:usedots,upriv,tm allow:lawrence,eugene,janice

Change the folder / names / users to your own ones obviously!
If its not going to be a time machine share, remove the “,tm”

Restart Netatalk

/etc/init.d/netatalk restart

You should be able to see the share in Time Machine Preferences.
See if you can backup. If you get a failure eg “Error 2”, make sure that the folder you use has write privileges for your user, then try again.

All in all pretty painless.

Proof it works –

Screen Shot 2013-04-01 at 1.43.08 PM

Flattr this!

I usually do most of my work these days on an Air. While its decent, it is rather dinky spacewise, and I tend to keep it empty of music so I have (barely) enough space for my dev stuff, and various embedded cross compile environments.

I still like listening to music though, so I keep that on secondary and even tertiary storage. This typically is one of the fine HP N36L / N40L NAS devices that I’m fond of buying. I have a few HP N36/N40L nas boxen in various places, some of which contain my music, but most of which contain backups, onsite, and offsite, as you can never be too careful!

I know this quite intimately, as the very machine I’m typing this on had a catastrophic SSD failure only a week ago.
Luckily I didn’t lose too much work!

Enough with the background, and back to the plot.

As I was saying, I needed some tunes to harass the staff with listen to at work.

I looked at a couple of solutions, and decided on forked-daapd, as that allegedly could share music to iTunes from the NAS music folder without too many headaches.

Most instructions were of the sort:

apt-get install forked-daapd

pico /etc/forked-daapd.conf

[edit music folder, save file]

/etc/init.d/forked-daapd restart

In my case it didn’t work.
It also didn’t really give much clue that it wasn’t working, and their website didn’t have much to go on.
I looked at compiling from scratch, but the guy making it uses Clang and Java stuff to build, and it just looked like too much hassle.

So, I had to troubleshoot even though I didn’t really want to spend the time.

My initial issue was something like the below:

forked-daapd wouldn’t run successfully (but it also didn’t complain, sigh). I could see that it wasn’t running on any port specified, and checking kern.log showed it was crapping out silently.

Mar 20 21:03:09 officenas kernel: [ 3392.026612] forked-daapd[9848] trap invalid opcode ip:7f8d8772958e sp:7f8d8176bf60 error:0 in[7f8d87722000+c000]

Running it in the foreground showed it was having issues creating the mDNS bits.

mdns: Failed to create service browser: Bad state

This eventually worked out to editing /etc/nsswitch.conf, adding the following to the host lines:

hosts: files mdns4_minimal dns mdns4

then restarting avahi

/etc/init.d/avahi-daemon restart

This got me past the bad state error, but then it was bombing out with a missing symbol avl_alloc_tree error.

I did an strace on the thing and found it was looking for libav under /lib vs under /var/lib

This was also documented here –, although sadly not until I found out myself.

Looks like the zfsonlinux is the culprit here, as that puts libav files in that folder. Tsk tsk.

I removed those libav files – rm /lib/libav* ****as I know what I’m doing**** (don’t randomly erase stuff unless you’re 3000% sure!), and sure enough, forked-daapd started up, and started blatting tons of output to the logs in a happy manner.
iTunes was also finally happy, and could see my music. Yay.

Took me about 2 hours to figure out sadly, but at least I sorted it out.

Hopefully this will save someone else the headache when they google for the error(s)!

Flattr this!

As my friends in high places have been talking about Ruby for a long long time now, I thought I might take a look at installing a Ruby based app on one of our servers. Sure, I could have hosted it on Heroku or similar (as I know people that know people), but I tend to do stuff in-house as China often decides to arbitrarily block useful 3rd party services at the drop of a hat.

Looked fairly simple I thought.

Bzzzt, wrong. (This is a bit of a diatribe, but hey, I have to whine somehow 😉 )

Seems Ruby has a little bit to go in terms of friendliness.

First up, was to follow the fairly simple instructions for installing the App I chose – (Kandan).

My first issue was this –

Installing eventmachine (0.12.10) with native extensions
Gem::Installer::ExtensionBuildError: ERROR: Failed to build gem native extension.

/usr/bin/ruby1.8 extconf.rb
extconf.rb:1:in `require': no such file to load -- mkmf (LoadError)
from extconf.rb:1

Gem files will remain installed in /var/lib/gems/1.8/gems/eventmachine-0.12.10 for inspection.
Results logged to /var/lib/gems/1.8/gems/eventmachine-0.12.10/ext/gem_make.out
An error occurred while installing eventmachine (0.12.10), and Bundler cannot continue.
Make sure that `gem install eventmachine -v '0.12.10'` succeeds before bundling.

Hmm, mkmf, whats that?

No idea, lets take a quick look at apt-cache.
Ok, so looks like we need ruby headers to compile.

A quick look google shows that at RubyForge shows that this has been an issue since oh, at least 2005.

Perhaps a nicer message might be – “Hey, I see you don’t have the ruby development headers installed, install some”, and maybe even download them.

Even Perl is more user friendly than that when it comes to missing libraries, and Perl is famous for being obscure.

Once I overcame that minor hurdle, the installer trundled away merrily, and failed on the next message

Gem::InstallError: cloudfuji_paperclip requires Ruby version >= 1.9.2.
An error occurred while installing cloudfuji_paperclip (3.0.3), and Bundler cannot continue.
Make sure that `gem install cloudfuji_paperclip -v '3.0.3'` succeeds before bundling.

Debian has Ruby 1.91, and Ruby 1.8 in stable.
*and* the previous compiled gem (eventmachine) required 1.8 specifically.
I’m already smelling versionitis…

[more wizened geeks will say apologetic things like:

#1 ah, but yes there is rb!
– Yes, but I’m coming at this from a fresh angle, and I don’t necessarily know about that.

#2 This is quite debian specific!
– Yes, but it is a rather major distro..

Lets see whats available from testing repo.

apt-cache search -t testing ^ruby | grep 1.9

ruby1.9.1 - Interpreter of object-oriented scripting language Ruby
ruby1.9.1-dev - Header files for compiling extension modules for the Ruby 1.9.1
ruby1.9.1-examples - Examples for Ruby 1.9
ruby1.9.1-full - Ruby 1.9.1 full installation
ruby1.9.3 - Interpreter of object-oriented scripting language Ruby, version 1.9.3

Ok, so 1.9.3 should do it, lets install that.

apt-get install -t testing ruby1.9.3

Re-run the Bundle installer, and…

Installing cloudfuji_paperclip (3.0.3)
Gem::InstallError: cloudfuji_paperclip requires Ruby version >= 1.9.2.
An error occurred while installing cloudfuji_paperclip (3.0.3), and Bundler cannot continue.
Make sure that `gem install cloudfuji_paperclip -v '3.0.3'` succeeds before bundling.

Lets double check.

> ruby --version
ruby 1.9.3p194 (2012-04-20 revision 35410) [x86_64-linux]

Lets see. Ruby 1.9.3 >= 1.9.2 in my math book, so wtf.

I can even install it fine manually, via “gem install cloudfuji_paperclip -v ‘3.0.3’” so its really full of poop.

I decide to take a different tack –
Looking at the gems folder though, I don’t see the gem libraries for 1.9.3 there, so I guess Ruby is full of crap again, and lying about the error, although then why does building the gem manually NOT fail. Sigh.

I decided to take the rvm route -> curl -L | bash -s stable –ruby


source /etc/profile.d/

(still need to add to apache www-data group, but first lets get this compiled)

rvm trundled away and installed 1.9.3 gems, so that *finally* cloudfuji_paperclip wasn’t bitching.

…and we get to the next error.

An error occurred while installing pg (0.12.2), and Bundler cannot continue.
Make sure that `gem install pg -v '0.12.2'` succeeds before bundling.

I run that manually, and

gem install pg -v '0.12.2'
Building native extensions. This could take a while...
ERROR: Error installing pg:
ERROR: Failed to build gem native extension.

/usr/local/rvm/rubies/ruby-1.9.3-p374/bin/ruby extconf.rb
checking for pg_config... no
No pg_config... trying anyway. If building fails, please try again with
checking for libpq-fe.h... no
Can't find the 'libpq-fe.h header

apt-get install libpq-dev solves that one.


bombing on sqlite.

apt-get install sqlite3


Its at this point I start thinking about puppet and how that does dependencies in a graceful manner, but I digress.

..and because I forget the development libraries, I need to also get those.

apt-get install libsqlite3-dev

Retry the bundle install, and *finally* getting a build.

Oh joy.

So, lets try run it.

bundle exec rake db:create db:migrate kandan:bootstrap

== CreateAttachments: migrating ==============================================
-- create_table(:attachments)
-> 0.0015s
== CreateAttachments: migrated (0.0016s) =====================================

== AddSessionsTable: migrating ===============================================
-- create_table(:sessions)
-> 0.0011s
-- add_index(:sessions, :session_id)
-> 0.0004s
-- add_index(:sessions, :updated_at)
-> 0.0004s
== AddSessionsTable: migrated (0.0020s) ======================================

== DeviseCreateUsers: migrating ==============================================
-- create_table(:users)
-> 0.0435s
-- add_index(:users, :email, {:unique=>true})
-> 0.0005s
-- add_index(:users, :ido_id, {:unique=>true})
-> 0.0005s
-- add_index(:users, :authentication_token, {:unique=>true})
-> 0.0005s
== DeviseCreateUsers: migrated (0.0452s) =====================================

== CreateChannels: migrating =================================================
-- create_table(:channels)
-> 0.0011s
== CreateChannels: migrated (0.0011s) ========================================

== CreateActivities: migrating ===============================================
-- create_table(:activities)
-> 0.0012s
== CreateActivities: migrated (0.0013s) ======================================

== AddGravatarHashToUsers: migrating =========================================
-- add_column(:users, :gravatar_hash, :text)
-> 0.0007s
== AddGravatarHashToUsers: migrated (0.0007s) ================================

== AddActiveToUsers: migrating ===============================================
-- add_column(:users, :active, :boolean, {:default=>true})
-> 0.0007s
== AddActiveToUsers: migrated (0.0007s) ======================================

Creating default user...
Creating default channel...
rake aborted!
undefined method `to_i' for #
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activemodel-3.2.11/lib/active_model/attribute_methods.rb:407:in `method_missing'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/attribute_methods.rb:149:in `method_missing'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/connection_adapters/column.rb:178:in `value_to_integer'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/connection_adapters/column.rb:78:in `type_cast'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/attribute_methods/dirty.rb:86:in `_field_changed?'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/attribute_methods/dirty.rb:63:in `write_attribute'
/usr/local/rvm/gems/ruby-1.9.3-p374/gems/activerecord-3.2.11/lib/active_record/attribute_methods/write.rb:14:in `channel_id='
/home/cubieboard/chat/kandan/lib/tasks/kandan.rake:35:in `block (3 levels) in '
/home/cubieboard/chat/kandan/lib/tasks/kandan.rake:23:in `each'
/home/cubieboard/chat/kandan/lib/tasks/kandan.rake:23:in `block (2 levels) in '
/usr/local/rvm/gems/ruby-1.9.3-p374/bin/ruby_noexec_wrapper:14:in `eval'
/usr/local/rvm/gems/ruby-1.9.3-p374/bin/ruby_noexec_wrapper:14:in `

Tasks: TOP => kandan:bootstrap
(See full trace by running task with --trace)

Oh look, *what* a suprise. Another error.
I’m starting to think that nobody actually tests this stuff in real life.

undefined method `to_i’

A bit of googling, and it looks like Ruby has changed functionality, and broken things in 3.2.3

See –

Plus, there are some security issues to in Rails (of course).

*and* Kandan guys have cancelled it.

However someone else has forked it, and is updating it.

So, lets wipe all of that, and start again, shall we.


cd ..
rm -r kandan –with-prejudice
git clone

edit the config/database.yaml

Add some sqlite3 (hey, the rest is, and at this point I just want something testable, I can tweak later).

adapter: sqlite3
host: localhost
database: db/production.sqlite3
pool: 5
timeout: 5000


exec rake db:create db:migrate kandan:bootstrap

Done. Yay.

Now to test.

bundle exec thin start
>> Using rack adapter
>> Thin web server (v1.3.1 codename Triple Espresso)
>> Maximum connections set to 1024
>> Listening on, CTRL+C to stop

Ok, now *thats* finally working, I just need to setup a apache_proxy to that port on the actual url it will be sitting on, and finally it should work.

*Famous last words*.

Way more painful than it really needed to be. Grr…

Flattr this!

WordPress has a nice feature called Custom Post types.

(An overview for those that code for WordPress is here – )

We typically use the Custom Post Types UI plugin ( )  to create, rather than write code.  The end result is the same though.


This allows us to create specifically enumerated sets of content relatively easily.

eg – if you were doing a Site for a Client, you may need a Portfolio section for their web.

With custom post types, you can setup a Portfolio post type, and add various taxonomies (meta data to group or organize these posts by later)

eg  Category


Eg if it was for a web design firm, you might add values into the  Category taxonomy like





Then choose those categories when you add a new Portfolio post so that you can later filter by any of those fields.

So far so good.


We also use another Plugin called QTranslate ( )

QTranslate allows us to add multiple translations for a post, or taxonomy in the WordPress UI.

Its pretty nifty, and a good tool for those that do multilingual wordpress sites.


QTranslate seems to be a little funky about how it works with Custom Post Types though – essentially its a little hit and miss whether it actually populates the backend with translation options.

After a bit of googling I found a nice solution to force it to parse taxonomies.

Add this into your Template functions.php at the end of the file, and QTranslate will work properly.

function qtranslate_edit_taxonomies(){
'public' => true ,
'_builtin' => false
$output = 'object'; // or objects
$operator = 'and'; // 'and' or 'or'

$taxonomies = get_taxonomies($args,$output,$operator);

if ($taxonomies) {
foreach ($taxonomies as $taxonomy ) {
add_action( $taxonomy->name.'_add_form', 'qtrans_modifyTermFormFor');
add_action( $taxonomy->name.'_edit_form', 'qtrans_modifyTermFormFor');
add_action('admin_init', 'qtranslate_edit_taxonomies');

Flattr this!

As we’re a pro-active sort of ISP, we often take a look at ways to improve things for clients under the hood.

One of those improvements, was our recent DMARC implementation for our mail servers.

DMARC is a newish standard which adds to our existing SPF setup, by allowing us to publish methods that tell other providers how mail from us should be processed. It also allows us to receive reports from other providers as to how they’re processing our mail.

Read more »