TechHobby

Recently a business where I had worked lost the use of one of its PCs due to corruption of the operating system. I was called in to help sort things out. My experiences made me think what might be a good checklist for any small business with one or two PCs, lets assume they are running windows.

A. Write a brief disaster recovery procedure for different scenarios, e.g theft, hardware damage, software corruption, accidental damage to data.
B. Have working/backup CDs of all software including operating systems and device drivers.
C. Have all codes and passwords recorded somewhere safe.
D. Have a backup of all critical data, both on site and off site, and instructions on how to restore it.

In considering this I have tried to be realistic rather than what you might do in an ideal world. The resources spent on precautions have to equate with the actual levels of risk. For example, daily imaging of every PC, storing data in carefully configured RAID arrays, is not for everyone, and might be wasted money and effort if the business cost of just re-loading some software and copying files from a memory stick backup is minimal.

Specific additional questions based upon what happened in practice are below:-

1. If the operating system won't boot: is there an Automated Recovery CD?, in this case there wasn't so Windows XP Pro had to be re-installed from scratch.

2. Are there working installation CDs for all critical software? In this case, one of the installation CDs turned out to be damaged, and that was the only copy.

3. Have all settings and passwords for internet services such as email been recorded in a readily available form?

All these precautions sound like common sense but may be more difficult to address than one might suppose. In a small business there may not be one person whose responsibility it is to see to all these things, ultimately it is the owner of the business who is responsible. However, he or she might not understand fully the implications of every technical issue, whilst IT contractors who have been brought in to do a specific task may not understand the implications of business issues. Responibility may have been delegated to a member of staff at one point, but then they leave, and a new person comes along, who, as it is a small business, may do a slightly different job to their predecessor because they have a different skill set. It is all to easy to assume that just because some action was taken with regards to a risk (i.e. buying a backup drive), that whatever has been put in place will actually work out when the worst happens and it comes to restoring everything. (Has the backup procedure been followed? Did the backup software work properly? Will the restore work on a new set up? How long were backups kept before being overwritten by the next backup?)

I don't want to be alarmist, but as data has become increasingly 'real' and 'tangible' as a business asset, it is probably worth while for anyone who works in a small business to ask some of these questions, because it might just save a lot of time and money when the PC won't start one not too distant Monday morning.

My P4 3Ghz XP Shuttle has found new life running puppy linux off of a memory stick, using a Vesa graphics driver, and connecting to the home network.

It was my main PC until it became increasingly unreliable and finally failed completely earlier this year, and I knew it was time to bite the dual-core + Vista bullet with a brand new PC. Since then I have played about with putting in different hard drives, optical drives and operating systems, trying to get the old PC going as a spare computer. However, the hard drive and optical drive kept disabling themselves on me, until I decided to let it run entirely from a memory stick, and after fiddling around, downloading various distributions, which did not work well for me, I remembered I had already installed Linux on a 2 Gb stick. Disabling the hard drive and optical drive in the BIOS, and inserting the stick in the back of the PC, allowed me boot into the pleasant little OS known as Puppy, browse the internet, word process, and do quite a lot of other things should I feel the need.

I recently inherited a Toshiba Portege pentium 3 laptop after buying my sister a new Acer dual core one running Vista for Christmas. The Windows 2000 operating system on the Toshiba was making it seem very slow, especially since the extra RAM slot was broken and hence it had a maximum of 64 megabytes. After installing Windows 98, the original operating system, it was working much better, and booted in about a minute rather than five minutes.

The other day I decided to install Linux on it as well, and luckily had an original set of Redhat 7.3 CDs to use. After partitioning the disk, the install took about forty-five minutes, but was then up and working, including X-Windows, without any problems whatsoever.

The reason I had wanted to have linux on that machine was so I could experiment with serial communications. This is hard to do on a modern PC that has only USB ports. Previously I had tried writing Windows 32-bit software to connect to a telnet server and relay the information to a terminal, with some limited success.

In Linux, I set up a process in the configuration files called 'agetty' which allowed a serial device to be used as an extra terminal. For my terminal I used an old Atari Mega ST computer with an original monochrome monitor, experimenting with various shareware and freeware terminal emulator programs. Once I had got the baud rate right, and was receiving recognisable text, I found that I could get complete emulation with a shareware program called CONNECT and partial emulation with other simpler programs, which was okay as long as I disabled colour in the Linux output, and steered clear of interactive text-user interface applications like PINE that would confuse a simple terminal.

The whole experience has made me appreciate better the complexities of what I previously saw as trivial: the correct coding, encoding, processing and display of text based user interaction.

I installed CentOS on my new Windows XP PC, having to resize the ntfs filesystem first to make space, with the intention of testing some open source postgres database replication software I was planning to use in my voluntary job. However, I stalled at ever doing this, maybe because there seemed to be too many variables outside my control.

Now with the space already partitioned, I decided to install a recent version of Suse Linux there instead. This went smoothly but it inevitably failed to recognise my USB wireless dongle. I eventually downloaded the right Ralink driver, and after installing the kernel source was able to compile it.

After some tinkering I was able to connect to my home wireless network. I added the DNS server address of the router in /etc/resolv.conf. To make it all load up at boot time I created a script in the /etc/init.d directory and inserted it to the startup sequence using 'chkconfig --add'.

#!/bin/sh
# BEGIN INIT INFO
# Provides: rc.local
# Required-Start: $network $syslog
# Required-Stop: $network $syslog
# Default-Start: 3 5
# Default-Stop: 0 1 2 6
# Description: whatever
# END INIT INFO
cd /root
cd rtl8187_linux_26.1025.0328.2007
./wlan0up
iwconfig wlan0 essid Belkin_Pre-N_123456
iwconfig wlan0 key 12345678909876543212345678
./wlan0dhcp

I am happy that I got it working, and it all seems quite robust and resposive.

For many months now, Real Player on my Apple Powerbook has been crashing out whenever I started to play any file or stream. This continued despite deleting the program and re-downloading and installing it.

Finally, I'm pretty sure that I've found out where the problem was - a corrupted cache. Trying to empty the cache lead to another crash, but turning it off altogether has solved the problem.

The cache does not seem to be important as my internet connection is very fast and reliable, and most of the streams I access are in the UK.

I have finally got a reasonable mobile data connection for my laptop (pictured below), through a tangental series of events.

For a while I've been using my bluetooth enabled Sony Ericson T-610 phone as a GSM modem for my Powerbook and my Palm T3, achieving the pitful speed of 9,600Kb/s. It came to my attention that I could connect a lot faster using GPRS (which is 2nd generation mobile technology). However, when I did this on my last holiday, I soon found that I was burning up credit on my Virgin Mobile pay as you go account very quickly. It was clear to me that a data-specific subscription was the only sane way to be using GPRS, but once I was back home, I soon forgot about it all, and was content to use GSM for the odd bit of browsing on my T3.

The issue re-occured when I was attending a local community-based group where people often had practical issues that needed resolving that required information which was freely available on the web, but they were non-tech savy and had no computer at home. Looking at the funding for the group, it stated that internet access would be provided, and yet it wasn't being. The group met in a public building with no technical facilities, so a mobile data connection seemed the obvious answer.

I found a cheap data card on ebay, and brought it in to try with the organisation's laptop. Using the SIM card from the group's mobile was unsuccesful as the contract did not include data. The SIM from my phone worked fine though, but the GSM connection was painfully slow. GPRS was still way too expensive an option using my pay-as-you-go SIM card, so a few weeks later I bought a subscription costing 25 pounds a month plus 7.50 a month for (virtually) unlimited data. This worked amazingly well in the laptop. I'd like to say the story ended there, but internal politics and lack of common sense intervened. The upper management who administered the group would not agree to have their laptop used for internet access. So I ended up taking back the card and SIM for my own personal use.

What I haven't mentioned is that the subscription came with a free Motorola phone, which is very smart, and I now use with my old Virgin mobile SIM card. It includes a built in camera that takes quite nice photos (see below).

The data card I had was an Option Globetrotter, that worked okay in my old Thinkpad laptop, but the software was very slow to load, being designed for an XP computer of reasonable spec. My Dad's newer Thinkpad worked a lot better with it. Even so, ideally I wanted to be able to go mobile on my Apple Powerbook laptop. To achieve this would need either a dedicated Mac card or some special software. As the former seemed to be non-existent, I opted for the latter, buying a  product called 'launch2net' from the Apple store. I was pleasantly surprised to find it recognised the card first go, but found that it did not seem able to connect. Then I found another cheap card on ebay that claimed to be Mac compatible. Although the bundled software rejected my SIM card as it was from another network, the launch2net utility recognised it and this time worked all the way. Finally my Powerbook has a decent mobile data connection that doesn't cost the earth.

Being quite against throwing away perfectly good computers and upgrading them for no good reason (other than to play a game with slightly more polygons) I finally felt it was time to retire my ~1GHz main PC and build a new one.

Why did I decide to build a PC rather than buy one ready made?

• First I wanted an Intel Mac, but it was taking too long to save up for.
• Macs only have specific hardware components, PCs in theory allow mix and match.
• Maybe I could save a little money.
• I had helped build systems before, but not that recently, so it would be a learning experience.

I checked the website of a very local dealer, and that lead me to a company in Manchester that sold bare-bones PCs, that is partly assembled units consisting of a case and motherboard. I bought a Shuttle XPC.

My first build was unsuccesful, even after changing the processor, memory and graphics card for ones on the manufacturers compatibility list. I sent the unit back, and the replacement worked a treat, until something went wrong while installing a parallel port. The retailer tested the system and said it was fine, which lead me to the conclusion that the graphics card had gone wrong, so I replaced it, with a better one, and lo and behold the system worked again.

The whole process took a month or more, and was a bit stressful at times, given that I did not have a lot of resources for problem solving faults and finding out which components worked or not. I am not sure I would buy a fully ready built system, but I might in future have the supplier build a system to my own specifications.

The final system has 2Gb of dual-channel memory, a Pentium 4 - 3.0GHz processor, a Sapphire 512Mb graphics card and 250Mb SATA hard drive and a standard DVD writer and I am happy now that it is all working.

Having networked a Mac and some PCs for a few years now, I have often run into difficulty when connecting from one to the other because of two problems:

1. the default workgroup name is 'workgroup', but my PCs are under 'home'

2. the dynamically allocated ip address of the Mac keeps changing, meaning that the required connection string one day will be different the next day

Finally I have made some headway, when I had a look at a configuration file on the Mac, stored in the Unix file structure. It is called:

/etc/smb.conf

Having experimented with Windows Networking under a few flavours of Linux in the past, I recognised that this was the 'Samba' configuration file, which sets up how the machine is viewed on a windows network. Within this file, I found two very useful options, which when changed, made things a lot easier:

workgroup = Home
netbios name = PowerBook

I also enabled home directory sharing.

[homes]
browseable=yes

Now I can simply refer to my Mac by its name, on my PCs, then enter in my Mac's username and password and I immediately have access to all my files.

This is a lot easier than going into the router configuration to find out my Mac's IP address, then having to type this in to connect.

I was asked to help set up a new web site where a content management system had been installed, and the old content was gradually being copied from the static pages into the database in a piecemeal fashion.

The trouble was, that to get to the old site, users had to click on a small link, which many of them missed, and hence could not get through to the info they wanted. The new site was incomplete, but becuase the C.M.S. has to be installed into the root of the web server, it was what users were getting, instead of the old site.

I took over the job of copying content from the old to the new site, but first I had to make sure that users would get to the old site, whilst I would get to the new site. This seemed hard, given that there was just one web address which had to be able to serve both the old pages to the public and the new pages to me whilst I was working on them.

My first thought was to copy the C.M.S. and database onto my Powerbook laptop, but I did not relish the sort of installation problems caused by different versions of PHP, and I also did not want to have to hack about with the configuration files to change the paths from test-server to production-server.

My second thought was to have a simple re-direct to the directory containing the old files. However, doing this broke the C.M.S. meaning I could not preview any of the pages I was working on. Finally, I hit upon the idea of scanning the web client's IP address, and based on this either redirecting to the old site, or continuing and processing the index page as per the C.M.S.'s internal workings.

The snippet of code I came up with was as follows:

	if ($_SERVER['REMOTE_ADDR']!='82.83.84.85')
	{
		header("Location: oldsite/index.htm");
		exit(0);
	}

...where 82.83.84.85 was the current dynamic-ish IP address assigned by my Internet Service Provider, which I assumed would not change very often.

I inserted this code into the beginning of index.php and there I had it, a web site that I could mess around with at will, without upsetting the all important web-site audience out in internet land.

A few weeks ago my old Thinkpad 600E met with an unfortunate accident, which I won't go into, and lost its LCD screen and CD-Rom drive.

For a few days, I wrote off the laptop totally as a lost cause. Then a brand new battery arrived in the post, which I had ordered before the laptop got broken. The old battery had a habit of getting down to 60% then running out of power.

Having spent this money on a new battery, and not wanting to have to send it back to Germany, where it was dispatched from, I decided to risk throwing good money after bad, and replace all the faulty parts. I opted to buy a new LCD panel, that is the display itself, not including the frame. Instead of getting another CD-Rom drive, I was lucky to find for the same price a DVD-Rom, CD-Recordable combo drive.

These parts only took a couple of days to arrive, but I waited for a Sunday afternoon when I was not busy to try and install them.

The LCD frame came apart with a little persuasion, and removing 6 screws, hidden behind little plastic circular pads. It was only connected in two places, a white plug for power, I think, and a thin flat ribbon cable for data. The new LCD display fitted snuggly in place, and much to my surprise, worked first go. The only difficulty in putting it back together was aligning the external brightness slider with its corresponding internal slider. I gave up on this and just put the frame back together, and by some miracle, in doing this the slider lined up and started working.

A much easier job was putting in the new DVD-Rom drive, which was simply a matter of pushing the catch and sliding the old drive out, and the new one in.

Finally, I put the new battery in, and set it charging for a few hours.

Thanks to an 'ahem' accident, the laptop now has a recordable drive, a new display (with only one hard to spot duff pixel), and a battery life of over two hours rather than 45 minutes. A saying that I will stick to in this case is: you never appreciate what you have, until you've broken it, and then had to fix it.