Skip to content

Firewall causes Rails update failure in NetBeans

If you are receiving the following error message when trying to update Rails in NetBeans 6.7:

ERROR: While executing gem … (Gem::RemoteFetcher::FetchError) SystemCallError: Unknown error – An established connection was aborted by the software in your host machine (http://gems.rubyforge.org/gems/actionmailer-2.3.3.gem)

…try momentarily disabling your firewall. In my case, Windows Firewall seemed to be torpedoing the process: after disabling it for a moment, the update went through perfectly.

Perhaps obvious, but also easily overlooked.

Credit: This post by “joseph george” reminded me to check the firewall settings.

Mazda MX-5 mk.1 spark plug lead change

Misfiring between 2000-3000 rpm suggested all was not well below the bonnet of my 1996 Mazda MX-5. On one journey, fortunately close to home, the misfire was so bad that it was reluctant to move above 3000 rpm at all, necessitating a gear change to force the revolutions up.

I had a new set of NGK BKR6E-11 spark plugs sitting around waiting to be fitted, but they didn’t seem to make any difference, so the next things to look at were the spark plug leads (also known as high tension leads). If the spark plug leads didn’t make a difference, the next thing to look at would be a coil pack; at over £200, this would be getting costly. So, at about £30, it made sense to try changing the ignition leads first. Again, MX5parts provided the spark plug leads at a reasonable price.

As usual, changing them was a simple task, but here are the photos anyway. If you need further guidance, you might want to consider buying the Veloce Mazda MX-5 1.8 enthusiast’s manual, or alternatively the 1.6i enthusiast’s manual, should that be more applicable.:

Spark plug leads ready
1. Spark plug leads ready
2. Old spark plug leads in the engine
2. Old spark plug leads in the engine
3. Remove electrical connectors; twisting them may be required
3. Remove electrical connectors; twisting them may be required
4. Grip tightly and remove each lead from engine
4. Grip tightly and remove each lead from engine
5. Reverse the process with the new leads
5. Reverse the process with the new leads
6. Step back and admire the new leads
6. Step back and admire the new leads

You can probably see from the last picture that the leads are a touch on the long side for the MX-5; I suppose that’s the result of buying budget leads.

As far as performance goes, however, the engine seems much happier. Not had any misfiring yet, and the engine is a lot happier at low revolutions, allowing me to stay in a lower gear. That, I believe, is having a positive effect on my fuel economy: by my calculations, since changing the leads and plugs, I managed over 30mpg from my 1.8i for the first time. And that’s without any motorway cruising, albeit a bit of dual carriageway driving.

Your mileage may vary, quite literally; let me know how you get on!

Mazda MX-5 mk.1 1.8i (1840cc) air filter change

When it comes to servicing one’s own car, the air filter is normally just about the easiest thing to change. Despite this, it’s often overlooked by lazy owners. Having a clogged air filter can increase fuel consumption and reduce power: it really makes sense to change it regularly. Fortunately, changing my Mazda MX-5 1.8i Merlot’s (also known as Mazda Miata in the U.S.A., and the Eunos Roadster in Japan) air filter was extremely straightforward.

My Veloce Mazda MX-5 1.8 enthusiast’s manual (you may want the 1.6i enthusiast’s manual instead) recommends inspection of the air filter every 6,000 miles. It should apparently be changed every 27,000 miles or 3 years. Armed with the workshop manual and a £12.77 brand new genuine Mazda air filter from MX5parts, I set to work.

Chances are that you could figure it out yourself, but on the off-chance that you’re somewhat timid or cagey about fiddling with your own car, here’s a step-by-step guide. (Usual disclaimers apply.) Apologies for the poor-quality camera-phone pictures.

1. Unplug the electrical connector
1. Unplug the electrical connector
2. Release the clamp
2. Release the clamp
3. Detach the trunking
3. Detach the trunking
4. Undo the four screws of the air filter casing
4. Undo the four screws of the air filter casing
5. Lift the casing up
5. Lift the casing up
6. Remove the air filter and any debris
6. Remove the air filter and any debris
7. Insert the new air filter
7. Insert the new air filter
8. Put screws, trunking, trunking clamp, electrical connector back
8. Put screws, trunking, trunking clamp, electrical connector back

That’s it! Drive off laughing like an idiot, enjoying the happier engine—and the consequently restored power.

Asus A8N-E nForce 4 Ultra RAID controller performance

With the recent HD video editing I’ve been doing, I’ve been on the performance hunt. My old Asus A8N-E setup had two Western Digital WD1200JS SATA II hard disk drives, in a RAID 0 configuration. I ran three tests, all straight after booting, and HDTach gave me these average readings:

Burst: 148.3 MB/s
Average read: 95.6 MB/s
Random access: 18.7s
CPU: 7%

This seemed low to me, considering that SATA II has a theoretical maximum of 3Gbps = 384 MB/s. I asked online about this but was told that this was normal.

I still wasn’t satisfied, though. I was suddenly struck by something I read – the Asus A8N-E has two onboard SATA controllers, one for ports 1 and 2, one for ports 3 and 4. Now both my hard drives were connected to SATA ports 1 and 2. What if I switched one drive to the other controller? I opened up the case and changed the configuration to use to ports 1 and 4. Restarted and ran HDTach, three times as before:

Burst: 276.6 MB/s
Average read: 106.0 MB/s
Random access: 18.7s
CPU: 5.3%

Yes, I couldn’t believe it either. Burst speed went up by over 83%! More importantly, the average read went up over 10% – a significant boost. CPU utilisation also appeared to decrease, but this could be normal variation.

So if you have an Asus A8N-E and want to optimise your RAID configuration – check your drives are making full use of both available controllers. Of course, usual caveats apply – make a backup, etc. RAID arrays have a nasty habit of exploding, slaying all.

NetBeans 4.1 with Windows 7 RC x64

Had some issues when I tried to install NetBeans 4.1 from the Open University CD-ROM, which is not surprising, considering its age and the determination of the Open University to keep on using legacy software.

First of all, the installer wouldn’t display correctly, with many graphical elements not appearing at all. Thanks to Simon over at abscforme.co.uk, I was able to get past this by forcing Windows 2000 compatibility mode.

Secondly, when going to File -> Open in NetBeans, Java crashed. This appeared to be a complete show-stopper. I even set up Windows XP Mode just to run it correctly. However, while I was doing that, I realised that the problem could lie with the Java Runtime: the Open University installer had installed the x86 version of the Java Runtime and the JDK. So I tried uninstalling the x86 versions and installing the Java SE Runtime Environment 5.0u19 and likewise the Java SE Development Kit 5.0u19 instead for Windows x64.

Upon installing NetBeans threw a paddy because it couldn’t find the JDK. I had to go to C:/Program Files (x86)/netbeans-4.1/etc/, set the permissions on the folder for Users to Write, and then edit the appropriate line in the netbeans.conf file to the following: netbeans_jdkhome=”C:\Program Files\Java\jdk1.5.0_19″. Then it was a case of putting the permissions on the folder back to as they were.

Upon restarting NetBeans, all seemed to be well…and so far, all is still well. As a bonus, using this version of the Java Runtime/JDK doesn’t force Windows 7 out of the Aero GUI. I’ll keep you posted if there are any further developments.

Having said all that, though…I really hope the Open University get a move on and switch to NetBeans 6.5 or later! It’s faintly ridiculous that courses that teach such a platform-independant technology as Java are so prescriptive of the platforms they will support.

Asus M3A76-CM doesn’t yet support ACC, despite its SB710 southbridge

It’s pretty annoying, but it appears that the Asus M3A76-CM motherboard doesn’t have an option in its BIOS to enable Advanced Clock Calibration. This is despite its southbridge supposedly supporting ACC.

Why is this such a big deal? Well, it isn’t for the average user, but if you’re an enthusiast and happen to have a dual-core CPU like the AMD Athlon X2 7750+ Black Edition, it certainly is. This CPU actually has two extra disabled cores, that have reputedly been enabled by flicking ACC on. You can imagine how useful this is if you happen to edit high-definition video, for example!

If anyone from Asus happens to be reading – small chance, I know – please do introduce this feature to the board in a BIOS update, if possible. I would imagine it would certainly boost the board’s popularity.

A Herefordshire welcome

To briefly explain what I’m doing with this blog – in short, it’s going to be a eclectic collection of thoughts, findings, and possibly even advice I’ve come across or come up with. Since I’m a geek, many will be computer or Internet based, but not exclusively: I’ll probably have the occasional thing to say about my Mazda MX-5 and my true sporting love, cricket.

Green fields; real ale; friendly folk – I could ask for a lot less from my surrounding county. Thanks for stopping by!